Electric power communication coupling intelligent power grid sensing, transmission and control integration method, device and system

Abstract

The invention relates to an electric power communication coupling intelligent power grid sensing, transmission and control integration method, device and system, and belongs to the technical field of electric power communication. Aiming at FDI attacks possibly existing in the system, a time sequence Kalman filtering method is adopted, the correlation between a consistency variable optimal estimation sequence and a measurement value sequence is compared, accurate perception of the FDI attacks is achieved, and the authenticity and reliability of consistency variable data are ensured. And secondly, in the aspect of communication control, when an FDI attack is sensed, an encryption interaction mode is switched into a quantum encryption mode based on a state-action evaluation function in a Q network, and respective quantum key distribution rates are selected and dynamically adjusted through cooperative work of a leading node and a following node. And finally, in the aspect of electrical control, constructing a consistency variable importance weight based on sequence correlation of encryption communication time delay and attack perception, performing normalization processing, and updating a consistency variable.

Description

Technical Field

[0001] The present invention relates to a method, device and system for integrating sensing, transmission and control of electric power communication coupling smart grid, belonging to the technical field of electric power communication. Background Art

[0002] The development of advanced information and communication technologies has significantly promoted the intelligentization and automation of smart grids. In smart grids, widely deployed IoT devices can collect key status information of energy storage equipment, distributed power sources, etc. in real time, and transmit this key information back to the control center through high-speed communication networks such as 5G and optical fiber, thereby supporting efficient real-time control of source-grid-load-storage.

[0003] However, the high coupling between the power system and the communication system in the smart grid makes the delay and security of data transmission directly affect the control performance of the power system. Especially in an untrusted environment, false data injection (FDI) attacks tamper with real measurement data and control instructions by injecting malicious data streams, affecting the convergence and accuracy of control decisions and endangering the balance of energy supply and demand. Therefore, how to effectively perceive FDI attacks and achieve low-latency and secure data transmission and precise electrical control is an important issue that needs to be solved urgently.

[0004] At present, there have been many studies on the integrated sensing, transmission and control methods of smart grids. The patent "Smart Grid Integrated Management System" proposes an integrated management system for smart grids, including: multiple sensor terminals, connected through a sensor network, used to obtain equipment or environmental information in each link of power transmission, substation, distribution and power consumption; a front-end application server, connected through the MIS bus, used to interact with the sensor network, and provide application services for each link itself and across links based on the equipment or environmental information of each link of power transmission, substation, distribution and power consumption; a management application platform, used to interact with the multiple front-end application servers, and provide integrated operations and data sharing of application services for each link itself and across links of power transmission, substation, distribution and power consumption. Based on the sensor network, a data sharing and integrated processing system for each link of power transmission, substation, distribution and power consumption is established to realize the management and monitoring of power resources, reduce the redundancy of work processes, and improve work efficiency. However, the above patent does not consider the impact of performance deviation from the perspective of power communication coupling, and does not consider the impact of data transmission delay and security on power grid control, causing false data and control instructions in the information domain to spread to the power domain through the communication network, causing power grid cascade oscillations and increasing control costs.

[0005] The patent "A Smart Grid Management System" proposes a smart grid management system, including: an information collection module, which is connected to the smart grid to collect power operation data; a power grid adjustment module, which is used to adjust the power transmission state of the power grid; a control module, which is respectively connected to the information collection module and the power grid adjustment module, to adjust the corresponding cutoff frequency of the filter when it is determined that the operation stability of the smart grid does not meet the requirements according to the variance of the voltage fluctuation amplitude of the transformer output line, or to preliminarily determine that the current monitoring accuracy of the transformer output line does not meet the requirements and secondarily determine the current monitoring accuracy of the transformer output line according to the average working current of the transformer output line; and, when the current monitoring accuracy is secondarily determined to be unsatisfactory, the current corresponding sampling period of the transformer output line is adjusted, thereby improving the operation stability of the smart grid. However, the above patent does not take into account the importance differences of network nodes, ignores the impact of communication delay and attack perception on consistency control, and does not use encryption interaction mode switching and elastic step size to adjust the importance weight, resulting in a slow convergence of the system or even failure to converge.

[0006] Therefore, there is an urgent need to design an integrated method, device and system for power communication-coupled smart grid sensing, transmission and control to effectively identify and defend against FDI attacks and achieve low-latency secure data transmission and precise electrical control. Summary of the invention

[0007] In order to solve the above technical problems, the purpose of the present invention is to provide a method, device and system for integrating sensing, transmission and control of power communication coupled smart grid.

[0008] The present invention provides a method for integrating sensing, transmission and control of a power communication-coupled smart grid, and the specific steps are as follows:

[0009] Firstly, an electrical control model based on a distributed consensus algorithm is proposed to reduce the dependence of control on the central node.

[0010] Secondly, in response to the FDI attack in the system, a communication control model integrating quantum encryption and classical encryption is proposed to dynamically adjust the encryption method and key distribution rate to ensure data security while reducing the transmission delay of the consistent variable of quantum encryption;

[0011] Next, a communication-electricity collaborative control model under FDI attack is proposed. Based on FDI attack perception, the encryption interaction mode and key distribution rate of the communication domain and the importance weight of the consistency variables of the power domain are dynamically adjusted to prevent the injected false data from spreading through the communication network and achieve fast and accurate convergence of consistency variables.

[0012] Based on the above three models, the time series Kalman filtering method is used to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of consistency variables and the measurement value sequence, ensuring the authenticity and reliability of consistency variable data;

[0013] In terms of communication control, when an FDI attack is detected, the system switches the encryption interaction mode to quantum encryption based on the state-action evaluation function in the Q network, and selects and dynamically adjusts the respective quantum key distribution rates through the collaborative work of the leader node and the follower node;

[0014] Finally, in terms of electrical control, the importance weights of consistency variables are constructed based on the serial correlation of encrypted communication delay and attack perception, and normalized, the consistency variables are updated, and the integrated optimization of power communication coupled smart grid sensing, transmission and control is carried out until the entire optimization cycle is completed and the optimal output power is output.

[0015] Furthermore, the electrical control model based on the distributed consistency algorithm is specifically:

[0016] Define the number of fusion terminals as J, and its set is At the jth fusion terminal s j The output control cost function F of the distributed generation in the controlled area j j (P j )for:

[0017]

[0018] In the formula, α j , γ j , β j They represent the quadratic coefficient, linear coefficient, and constant coefficient of the distributed generation output cost function in the area j, respectively. j is the output power of distributed power sources in area j;

[0019] Based on the Lagrange multiplier method and KKT conditions, the consistency variable λ is introduced j , indicating that at the jth fusion terminal s j The Lagrange multiplier of the controlled area; the definition of the consistency control process can be divided into K iterations, and the consistency variable λ of the jth fusion terminal j (k+1), the update process at the kth iteration is expressed as:

[0020]

[0021] In the formula, ξ j (k) represents the number of λ in the kth iteration j (k) is the update step size; τ j′,j (k) is the number of s in the kth iteration j′to j The communication delay of j′,j (k) is the kth iteration s j′ to j The importance weight of the consistency variable; j′ (k-τ j′,j (k)) represents the consideration of communication delay τ j′,j (k) Next, s j From j′ Received consistency variable; j-th fusion terminal s j The consistency variable λ of the controlled area j The convergence value of the optimal distributed generation output power P corresponding to the area j j , expressed as

[0022]

[0023] Furthermore, the communication control model integrating quantum encryption and classical encryption is specifically as follows:

[0024] definition is the encrypted interaction mode between fusion terminals; j,j′ (k) = 1 means s j With s j′ Quantum encryption is used to interact with each other; j,j′ (k) = 0 means that the classic encryption method is used for interaction; definition and are the quantum node set and the classical node set respectively; define τ j,j′ (k) is the number of s in the kth iteration j With s j′ The encrypted communication delay between is expressed as:

[0025]

[0026] In the formula, and Respectively represent the kth iteration s j With s j′ The communication delay between classical encryption and quantum encryption.

[0027] Furthermore, the communication-electrical collaborative control model under the FDI attack is specifically as follows:

[0028] During the consistency control process, the FDI attacker intercepts s j With s j′ The classical channel data between them is injected with false data to destroy the convergence of the consistency variables and reduce the control performance; define is the number of times the attacker is in s in the kth iteration j With sj′ The false data injected during the tampering is expressed as:

[0029]

[0030] In the formula, Θ j′,j (k) represents the kth iteration of the attacker in s j With s j′ Indicator variable for inter-FDI attacks, Θ j′,j (k) = 1 means s j With s j′ The communication data between them is attacked by FDI, otherwise Θ j′,j (k) = 0;

[0031] Communication-electrical cooperative control model under FDI attack, consistency variable λ of the jth fusion terminal j (k+1), the update process at the kth iteration is expressed as:

[0032]

[0033] Furthermore, the specific method of using the time series Kalman filtering method to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of the consistency variable and the measurement value sequence is as follows:

[0034] Construct s in the kth iteration j With s j′ The optimal estimate of the consistency variable between time series X j′,j (k) = [X j′,j (kk * ),…,X j′,j (k-1),X j′,j (k)] and the measurement value time series Y j′,j (k) = [Y j′,j (kk * ),…,Y j′,j (k-1),Y j′,j (k)]; FDI attack perception is performed by comparing the correlation of the two sequences and comparing them with the threshold; the cosine similarity is used to measure the correlation between the optimal estimation value sequence and the measurement value sequence; in the kth iteration, s j With s j The correlation between ′ j′,j (k) and FDI attack perception are expressed as:

[0035]

[0036]

[0037] In the formula, ||.|| represents the Euclidean norm; X j′,j (k) = [X j′,j (kk * ),…,X j′,j (k-1),X j′,j (k)] represents the optimal estimate sequence, whose elements include the (kk * ) iterations to the optimal estimate of the kth iteration; Y j′,j (k) = [Y j′,j (kk * ),…,Y j′,j (k-1),Y j′,j (k)] represents the measurement value sequence; ΔF represents the FDI attack perception threshold; Indicates that s in the kth iteration j With s j′ The indicator variable of FDI attack perception between Indicates the perception of j With s j′ There is an FDI attack; conversely,

[0038] Furthermore, the system switches the encryption interaction mode to the quantum encryption mode based on the state-action evaluation function in the Q network, and through the collaborative work of the leader node and the follower node, the specific method of selecting and dynamically adjusting the respective quantum key distribution rates is:

[0039] The fusion terminal that directly interacts with the distribution network control center is defined as the leader node, and the set is represented as Other terminals that cannot directly interact with each other are follower nodes, and the set is represented as To ensure consistent convergence, the leader node and the quantum nodes with which it has communication connections use quantum encryption interaction;

[0040] Follow Nodes j Based on the Q network, construct the state-action evaluation function Q(Φ j (k),s j′ ,y j′,j (k)); Q(Φ j (k),s j′ ,y j′,j (k)=1) means that in state Φ j (k) Next j With s j′ The value of quantum encryption interaction between j (k),s j′ ,y j′,j (k) = 0) represents the value of using the classic encryption interaction method; therefore, the encryption interaction method decision is expressed as:

[0041] y j′,j (k) = Ξ{Q(Φ j (k),s j′ ,1)>Q(Φ j (k),s j′ ,0)} (9);

[0042] Based on the hierarchical collaborative DQN algorithm, the quantum key distribution rate is dynamically adjusted according to the interaction requirements, the node's own distribution rate and upper limit, and the key inventory. First, the leader node selects the key distribution rate based on the key distribution value function in the Q network. Second, the follower node makes further adjustments based on the leader node's adjusted distribution rate and its own network status to ensure that the distribution rate is consistent with the neighboring nodes in the interaction. Finally, multiple iterations and feedback optimization are used to gradually approach the optimal strategy.

[0043] The leader node s of the kth iteration j With follower nodes s j′ The initial key distribution rate a between j,j′ (k) is expressed as:

[0044]

[0045] In the formula, Distribute value functions for keys; Represents the leader node s j To all adjacent follower nodes s j′ The sum of the values ​​corresponding to the key distribution;

[0046] Specify the node s to follow in the kth iteration j′ The key distribution rate It is expressed as:

[0047]

[0048] In the formula, for The distribution rate and upper limit of Representation and There is an interactive leader node; the first part in brackets is the remaining available distribution rate; in the numerator, express In Status Next, select j′ The Q value corresponding to the key distribution; the denominator is The sum of the Q values ​​of other follower nodes that interact;

[0049] Construct a loss function to evaluate the joint optimization performance of encryption interaction mode and key distribution rate; define the kth iteration leader node s jThe loss variable is R j (k), to evaluate the number of encryption failures caused by key resource waste and insufficient quantum keys, and to construct the loss function L j for:

[0050]

[0051] Where K is the number of iterations at convergence, γ j is a discount factor used to evaluate the impact of future rewards on current decisions; Indicates that in the k+1th iteration, the leader node s j In the state Φ j Select the best follower node under (k+1) The maximum Q value corresponding to key distribution;

[0052] After each iteration, the evaluation results of the loss function are fed back to each node to update its Q value and quantum key distribution rate strategy. Through continuous iteration and optimization, dynamic adjustment of the quantum key distribution rate strategy is achieved.

[0053] Furthermore, based on the serial correlation of encrypted communication delay and attack perception, the importance weight of consistency variables is constructed, and normalized, the consistency variables are updated, and the integrated optimization of power communication coupling smart grid sensing, transmission and control is performed until the entire optimization cycle is completed. The specific method of outputting the optimal output power is:

[0054] An electrical consistency control optimization method considering communication delay and attack perception is proposed. After detecting an FDI attack, the consistency variable importance weight based on the encryption communication delay and sequence correlation is constructed according to the delay of the received message and the encryption interaction mode. The encryption interaction mode switching and elastic step size are used to prevent false signals from spreading to the global network through the communication bridge and causing cascading oscillations.

[0055] Considering the encryption communication delay and attack perception, construct the kth iteration s j With s j′ The importance weight of the consistent variables between is expressed as:

[0056]

[0057] Where Δ(k) is the elastic step size of the kth iteration, which gradually approaches 0 as the number of iterations increases, ensuring that the adjustment amplitude is large in the initial stage of the algorithm and becomes smaller in the later stage of the iteration to ensure consistent convergence; φ τ With φ F are adjustment factors, which are used to adjust the impact of encryption communication delay and sequence correlation on importance weight;

[0058] The weights are normalized and expressed as:

[0059]

[0060] When j′ to j When there is an FDI attack, the consistency variables detected during the FDI attack iteration will be discarded to ensure that the update of the consistency variables will not be affected by false data. j The consistency variable λ j (k+1), the update process at the kth iteration is expressed as:

[0061]

[0062] An integrated device for sensing, transmission and control of electric power communication coupling smart grid, comprising:

[0063] Modeling unit: used to first propose an electrical control model based on a distributed consensus algorithm to reduce the dependence of regulation on the central node; secondly, in response to the FDI attack in the system, a communication control model that integrates quantum encryption and classical encryption is proposed to dynamically adjust the encryption method and key distribution rate to ensure data security while reducing the transmission delay of the consistency variable of quantum encryption; then, a communication-electrical collaborative control model under FDI attack is proposed, which dynamically adjusts the encryption interaction method and key distribution rate of the communication domain and the importance weight of the consistency variable in the power domain based on FDI attack perception to avoid the spread of injected false data through the communication network and achieve rapid and accurate convergence of consistency variables;

[0064] FDI attack perception unit based on time series Kalman filtering: Based on the above three models, the time series Kalman filtering method is used to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of the consistency variable and the measurement value sequence, ensuring the authenticity and reliability of the consistency variable data;

[0065] Joint optimization unit of encryption interaction mode and quantum key distribution rate based on hierarchical collaborative DQN: used in communication control. When an FDI attack is detected, the system switches the encryption interaction mode to quantum encryption mode based on the state-action evaluation function in the Q network, and selects and dynamically adjusts the respective quantum key distribution rates through the collaborative work of the leader node and the follower node;

[0066] Electrical consistency control optimization unit considering communication delay and attack perception: In terms of electrical control, it is used to construct consistency variable importance weights based on the serial correlation of encrypted communication delay and attack perception, and perform normalization, update consistency variables, and perform integrated optimization of power communication coupled smart grid sensing, transmission and control until the end of the entire optimization cycle and output the optimal output power.

[0067] Furthermore, the FDI attack perception unit based on temporal Kalman filtering includes:

[0068] Estimated value time series calculation module: uses the Kalman filtering method to calculate the estimated value time series of the consistency variable and transmits the result to the correlation calculation module;

[0069] Measurement value time series calculation module: uses Kalman filtering method to calculate the consistency variable measurement value time series, and transmits the result to the correlation calculation module;

[0070] Correlation calculation module: uses cosine similarity to measure the correlation between the optimal estimation value sequence and the measurement value sequence, and transmits the result to the FDI attack perception module;

[0071] FDI attack perception module: FDI attack perception is performed by comparing the correlation between two sequences and comparing them with the threshold, and the results are transmitted to the consistency variable discarding module.

[0072] Furthermore, the encryption interaction mode based on hierarchical collaborative DQN and quantum key distribution rate joint optimization unit includes:

[0073] DQN network module: Based on the loss function calculation results, perform intelligent agent learning, update the Q value and quantum key distribution rate strategy. Based on the current state information, calculate the state-action evaluation function and transmit the result to the encryption interaction mode decision module, the leader node key distribution rate decision module and the follower node key distribution rate decision module;

[0074] Encryption interaction mode decision module: Based on the DQN network module, it calculates and compares the state-action evaluation function size, makes node encryption interaction mode decisions, and transmits the results to the loss function calculation module and the electrical consistency control optimization unit that considers communication delay and attack perception;

[0075] Leader node key distribution rate decision module: Based on the DQN network module, the leader node key distribution rate is determined according to the key distribution value function in the Q network, and the result is transmitted to the follower node key distribution rate decision module, the loss function calculation module and the electrical consistency control optimization unit considering communication delay and attack perception;

[0076] Follower node key distribution rate decision module: Based on the DQN network module, it determines the follower node key distribution rate according to the leader node's distribution rate and its own network status, and transmits the result to the loss function calculation module and the electrical consistency control optimization unit considering communication delay and attack perception;

[0077] Loss function calculation module: The loss function is calculated based on the number of encryption failures caused by key resource waste and insufficient quantum keys. After each iteration, the evaluation result of the loss function is fed back to the DQN network module to update its Q value and quantum key distribution rate strategy. Through continuous iteration and optimization, dynamic adjustment of the quantum key distribution rate strategy is achieved.

[0078] Furthermore, the electrical consistency control optimization unit considering communication delay and attack perception includes:

[0079] Importance weight calculation module: calculates the importance weight of consistency variables based on encryption communication delay and sequence correlation, normalizes the importance weight, and transmits the result to the consistency variable update module based on importance weight and attack perception;

[0080] Consistency variable discarding module: based on the FDI attack perception results, the consistency variables detected by FDI attack are discarded to ensure that the update of consistency variables will not be affected by false data, and the results are transmitted to the consistency variable updating module based on importance weight and attack perception;

[0081] Consistency variable update module based on importance weight and attack perception: updates the consistency variables of each node according to the consistency variable importance weight and the consistency variable discarding result of FDI attack perception, and transmits the result to the output power calculation module;

[0082] Output power calculation module: Calculates the output power of each node based on the update results of the consistency variables.

[0083] An integrated system of electric power communication coupling smart grid sensing, transmission and control, including an electrical control layer, a classical channel transmission layer, a quantum key distribution layer, and a communication control layer;

[0084] The electrical control layer includes multiple distribution substations, each of which is optimized by a fusion terminal for electrical control decisions, and controls the output of distributed power sources to meet the energy supply and demand balance;

[0085] The classical channel transmission layer is composed of 5G channels and is used to transmit data encrypted by quantum keys or classical keys;

[0086] The quantum key distribution layer is composed of quantum channels and is used to distribute quantum keys.

[0087] By means of the above scheme, the present invention has at least the following advantages:

[0088] 1. The present invention proposes an integrated control architecture for sensing, transmission and control of smart grids. First, taking into account the complexity and dynamics of the power grid, an electrical control model based on a distributed consistency algorithm is proposed to reduce the dependence of control on central nodes. Secondly, in response to possible FDI attacks in the system, a communication control model that integrates quantum encryption and classical encryption is proposed to dynamically adjust the encryption method and key distribution rate to ensure data security while reducing the transmission delay of consistency variables in quantum encryption. Finally, a communication-electrical collaborative control model under FDI attacks is proposed. Based on FDI attack perception, the encryption interaction method and key distribution rate in the communication domain are dynamically adjusted, as well as the importance weights of consistency variables in the power domain to prevent the injected false data from spreading through the communication network and achieve rapid and accurate convergence of consistency variables.

[0089] 2. The present invention proposes an integrated optimization algorithm for sensing, transmission and control of smart grids based on communication-power domain performance deviation. First, in view of the possible FDI attacks in the system, the time series Kalman filtering method is adopted to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of consistency variables and the measurement value sequence, thereby ensuring the authenticity and reliability of consistency variable data. Secondly, in terms of communication control, when an FDI attack is perceived, the encryption interaction mode is switched to the quantum encryption mode based on the state-action evaluation function in the Q network, and the respective quantum key distribution rates are selected and dynamically adjusted through the collaborative work of the leader node and the follower node. Finally, in terms of electrical control, the importance weights of the consistency variables are constructed based on the encryption communication delay and the serial correlation of attack perception, and normalized processing is performed to update the consistency variables.

[0090] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show a certain embodiment of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0092] Figure 1 It is a structural schematic diagram of the power communication coupled smart grid sensing, transmission and control integrated system of the present invention;

[0093] Figure 2 It is a flow chart of the integrated method of sensing, transmission and control of power communication coupled smart grid of the present invention;

[0094] Figure 3 It is a structural schematic diagram of the power communication coupled smart grid sensing, transmission and control integrated device of the present invention. DETAILED DESCRIPTION

[0095] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0096] The present invention includes three parts: power communication coupled smart grid sensing, transmission and control integrated system, method and device, which are specifically described as follows.

[0097] The present invention proposes a power communication coupled smart grid sensing, transmission and control integrated system such as Figure 1 As shown, it includes an electrical control layer, a classical channel transmission layer, a quantum key distribution layer, and a communication control layer. The electrical control layer contains multiple distribution substations. Each substation optimizes the electrical control decision by the fusion terminal, and satisfies the energy supply and demand balance by controlling the output of distributed power sources. The classical channel transmission layer is composed of 5G channels, which are used to transmit quantum keys or data encrypted by classical keys. The quantum key distribution layer is composed of quantum channels, which are used to distribute quantum keys. Due to the high cost of quantum communication, only some fusion terminals support quantum communication, which are called quantum nodes, and other terminals are called classical nodes. The fusion terminal optimizes the output of distributed power sources based on consistency control. Specifically, each terminal interacts with adjacent terminals to iteratively converge the consistency state variables and adjust the power output according to the power deviation.

[0098] The integrated sensing, transmission and control system of power communication coupled smart grid realizes the integration of sensing, transmission and control of smart grid. In terms of perception, the fusion terminal performs FDI attack perception and attack location based on the consistency variables received from the adjacent terminals. In terms of transmission, based on the FDI attack perception results, the encryption interaction mode and quantum key distribution rate are dynamically adjusted to reduce the transmission delay of the consistency variables of quantum encryption. In terms of electrical control, based on the encryption communication delay and attack perception results, the importance weight of the consistency variable is adjusted to accelerate the convergence speed.

[0099] The present invention proposes a method for integrating power communication coupling with smart grid sensing, transmission and control. The specific process is as follows: Figure 2 shown.

[0100] S1 Smart Grid Sensing, Transmission and Control Integrated Control Architecture

[0101] The present invention proposes an integrated control architecture for sensing, transmission and control of smart grids. First, considering the complexity and dynamics of the power grid, an electrical control model based on a distributed consistency algorithm is proposed to reduce the dependence of control on central nodes. Secondly, in response to possible FDI attacks in the system, a communication control model that integrates quantum encryption and classical encryption is proposed to dynamically adjust the encryption method and key distribution rate to ensure data security while reducing the transmission delay of consistency variables in quantum encryption. Finally, a communication-electrical collaborative control model under FDI attacks is proposed. Based on FDI attack perception, the encryption interaction method and key distribution rate of the communication domain and the importance weight of the consistency variables in the power domain are dynamically adjusted to avoid the spread of injected false data through the communication network and achieve rapid and accurate convergence of consistency variables.

[0102] S1.1 Electrical control model based on distributed consensus algorithm

[0103] Assume that the number of fusion terminals is J, and its set is At the jth fusion terminal s j The output control cost function F of the distributed generation in the controlled area j j (P j ) is expressed as

[0104]

[0105] In the formula, α j , γ j , β j They represent the quadratic coefficient, linear coefficient, and constant coefficient of the distributed generation output cost function in the area j, respectively. j is the output power of distributed power in the area j. In order to realize the economic control of distributed power, the optimization goal is to minimize the total control cost of distributed power.

[0106] Based on the Lagrange multiplier method and KKT conditions, the consistency variable λ is introduced j , indicating that at the jth fusion terminal s j The Lagrange multiplier of the controlled region is adjusted by j Explore the optimal output power control scheme. Considering the complexity and dynamics of the power grid, the present invention adopts distributed control to adjust λ j Specifically, each terminal exchanges information with its neighboring terminals and uses a subgradient method to adjust the consistency variable λ. j Iterate and update, and adjust the power output according to the power deviation. Assume that the consistency control process can be divided into K iterations, and the consistency variable λ of the jth fusion terminal j (k+1), the update process at the kth iteration is expressed as

[0107]

[0108] In the formula, ξ j (k) represents the number of λ in the kth iteration j (k) is the update step size. j′,j (k) is the number of s in the kth iteration j′ to j The communication delay of j′,j (k) is the kth iteration s j′ to j The importance weight of the consistency variable reflects the contribution of the consistency variable to the update of the consistency variable. j′ (k-τ j′,j (k)) represents the consideration of communication delay τ j′,j (k) Next, s j From j′ Received consistency variable. The jth fusion terminal s j The consistency variable λ of the controlled area j The convergence value of the optimal distributed generation output power P corresponding to the area j j , expressed as

[0109]

[0110] S1.2 Communication control model integrating quantum encryption and classical encryption

[0111] In order to achieve FDI attack protection, quantum encryption and classical encryption are integrated. Definition is the encrypted interaction mode between fusion terminals. j,j′ (k) = 1 means s j With s j′ Quantum encryption is used to interact with each other; j,j′ (k) = 0 means that the classic encryption method is used for interaction. Definition and are the quantum node set and the classical node set respectively. Define τ j,j′ (k) is the number of s in the kth iteration j With s j′ The encrypted communication delay between

[0112]

[0113] In the formula, and Respectively represent the kth iteration s j With s j′ The communication delay between classical encryption and quantum encryption.

[0114] S1.3 Communication-electrical collaborative control model under FDI attack

[0115] During the consistency control process, the FDI attacker intercepts s j With s j′ The classical channel data between the two channels is used to inject false data to destroy the convergence of the consistency variables and reduce the control performance. is the number of times the attacker is in s in the kth iteration j With s j′ The false data injected into the time, the consistency variable after tampering is expressed as

[0116]

[0117] In the formula, Θ j′,j (k) represents the kth iteration of the attacker in s j With s j′ Indicator variable for inter-FDI attacks, Θ j′,j (k) = 1 means s j With s j′ The communication data between them is attacked by FDI, otherwise Θ j′,j (k)=0.

[0118] Considering the impact of FDI attack and encryption delay on consistency control, a communication-electrical cooperative control model under FDI attack is proposed. The consistency variable λ of the jth fusion terminal j (k+1), the update process at the kth iteration is expressed as

[0119]

[0120] S2 Smart grid sensing, transmission and control integrated optimization algorithm based on communication-power domain performance deviation

[0121] The present invention proposes an integrated optimization algorithm for sensing, transmission and control of smart grids based on communication-power domain performance deviation. First, in response to possible FDI attacks in the system, the present invention adopts a time series Kalman filtering method to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of consistency variables and the measurement value sequence, thereby ensuring the authenticity and reliability of consistency variable data. Secondly, in terms of communication control, when an FDI attack is perceived, the system switches the encryption interaction mode to the quantum encryption mode based on the state-action evaluation function in the Q network, and selects and dynamically adjusts the respective quantum key distribution rates through the collaborative work of the leader node and the follower node. Finally, in terms of electrical control, the importance weights of the consistency variables are constructed based on the encryption communication delay and the serial correlation of attack perception, and normalized processing is performed to update the consistency variables.

[0122] S2.1 FDI attack perception based on temporal Kalman filtering

[0123] Kalman filtering is a linear unbiased minimum mean square error estimation method that can iteratively estimate the measurement value of the consistency variable and perceive FDI attacks based on the similarity between the estimated value time series and the measured value time series. Specifically, when there is no FDI attack, the estimated value sequence of the consistency variable should be highly correlated with the measured value sequence. However, FDI attacks change the measured value by injecting false data, making it inconsistent with the estimated value. The consistency variable is updated sequentially and the attack location is achieved based on attack perception.

[0124] Construct s in the kth iteration j With s j′ The optimal estimate of the consistency variable between time series X j′,j (k) = [X j′,j (kk * ),…,X j′,j (k-1),X j′,j (k)] and the measurement value time series Y j′,j (k) = [Y j′,j (kk * ),…,Y j′,j (k-1),Y j′,j (k)] FDI attack perception is performed by comparing the correlation of the two sequences and comparing them with the threshold. The cosine similarity is used to measure the correlation between the optimal estimation value sequence and the measurement value sequence. j With s j′ The correlation between j′,j (k) and FDI attack perception are expressed as

[0125]

[0126]

[0127] In the formula, ||.|| represents the Euclidean norm. j′,j (k) = [X j′,j (kk * ),…,X j′,j (k-1),X j′,j (k)] represents the optimal estimate sequence, whose elements include the (kk * ) iterations to the optimal estimate of the kth iteration. j′,j (k) = [Y j′,j (kk * ),…,Y j′,j (k-1),Y j′,j (k)] represents the measurement value sequence. ΔF represents the FDI attack perception threshold. Indicates that s in the kth iterationj With s j′ The indicator variable of FDI attack perception between Indicates the perception of j With s j′ There is an FDI attack between. S2.2 Joint optimization of encryption interaction mode and quantum key distribution rate based on hierarchical collaborative DQN

[0128] The present invention proposes a joint optimization method of encryption interaction mode and quantum key distribution rate based on hierarchical collaborative DQN. First, the nodes are divided into leading nodes and follower nodes according to the importance of their positions, and the encryption interaction mode decision is made based on the state-action evaluation function in the Q network. Secondly, through the collaboration of the leading node and the follower node, the quantum key distribution rate of nodes at all levels is formulated and dynamically adjusted. Through the mutual collaboration of Q networks at all levels, the proposed method gives priority to ensuring that the leading node has a sufficient number of keys to correctly and unbiasedly transmit key control information to the follower nodes, avoids the large-scale spread of false data from the leading node, and improves the accuracy of consistency convergence.

[0129] S2.2.1 Optimization of encryption interaction method

[0130] Each quantum node needs to dynamically adjust the encryption interaction mode according to the consistency variable gap and FDI attack perception. The fusion terminal that directly interacts with the distribution network control center is defined as the leader node, and the set is represented as Other terminals that cannot directly interact with each other are follower nodes, and the set is represented as To ensure consistent convergence, the leader node and the quantum nodes with which it has communication connections use quantum encryption interaction.

[0131] Follow Nodes j Based on the Q network, construct the state-action evaluation function Q(Φ j (k),s j′ ,y j′,j (k)). Q(Φ j (k),s j′ ,y j′,j (k)=1) means that in state Φ j (k) Next j With s j′ The value of quantum encryption interaction between j (k),s j′ ,y j′,j (k) = 0) represents the value of using the classic encryption interaction method. Therefore, the encryption interaction method decision is expressed as

[0132] y j′,j (k) = Ξ{Q(Φ j(k),s j′ ,1)>Q(Φ j (k),s j′ ,0)} (9)

[0133] S2.2.2 Quantum key distribution rate optimization

[0134] The proposed hierarchical collaborative DQN algorithm dynamically adjusts the quantum key distribution rate according to the interaction requirements, the node's own distribution rate and upper limit, and the key inventory. First, the leader node selects the key distribution rate based on the key distribution value function in the Q network. Secondly, the follower node makes further adjustments based on the leader node's adjusted distribution rate and its own network status to ensure that the distribution rate is consistent with the neighboring nodes in the interaction. Finally, multiple iterations and feedback optimization gradually approach the optimal strategy.

[0135] The leader node s of the kth iteration j With follower nodes s j′ The initial key distribution rate a between j,j′ (k) is expressed as

[0136]

[0137] In the formula, Distribute value functions for keys. Represents the leader node s j To all adjacent follower nodes s j′ The sum of the values ​​corresponding to key distribution.

[0138] Specify the node s to follow in the kth iteration j′ The key distribution rate Expressed as

[0139]

[0140] In the formula, for The distribution rate and upper limit of Representation and There is an interactive leader node. The first part in brackets is the remaining available distribution rate. In the numerator, express In Status Next, select j′ The denominator is the same as The sum of the Q values ​​of other follower nodes that interact.

[0141] Construct a loss function to evaluate the joint optimization performance of encryption interaction mode and key distribution rate. Define the k-th iteration leader node s j The loss variable is R j(k), to evaluate the number of encryption failures caused by key resource waste and insufficient quantum keys, and to construct the loss function L j for

[0142]

[0143] Where K is the number of iterations at convergence, γ j is a discount factor used to evaluate the impact of future rewards on current decisions. Indicates that in the k+1th iteration, the leader node s j In the state Φ j Select the best follower node under (k+1) The maximum Q value corresponding to key distribution.

[0144] After each iteration, the evaluation result of the loss function is fed back to each node to update its Q value and quantum key distribution rate strategy. Through continuous iteration and optimization, the quantum key distribution rate strategy can be dynamically adjusted.

[0145] S2.3 Electrical consistency control optimization considering communication delay and attack perception

[0146] Traditional consistency control methods usually assume that the importance weights of all nodes in the network are the same, without considering the importance and state differences of the nodes, which slows down the convergence of the system or even prevents it from converging. To address this problem, an electrical consistency control optimization method considering communication delay and attack perception is proposed. After detecting an FDI attack, the importance weights of consistency variables based on encrypted communication delay and sequence correlation are constructed according to the delay of the received message and the encryption interaction mode. The encryption interaction mode switching and elastic step size are used to prevent false signals from spreading to the global network through the communication bridge and causing cascading oscillations.

[0147] Considering the encryption communication delay and attack perception, construct the kth iteration s j With s j′ The importance weight of the consistent variables between

[0148]

[0149] Where Δ(k) is the elastic step size of the kth iteration, which gradually approaches 0 as the number of iterations increases, ensuring that the adjustment amplitude is large in the initial stage of the algorithm and becomes smaller in the later stages of the iteration to ensure consistent convergence. τ and φ are adjustment factors, which are used to adjust the impact of encryption communication delay and sequence correlation on the importance weight.

[0150] The weights are normalized and expressed as

[0151]

[0152] When j′ to j When there is an FDI attack, the consistency variables detected during the FDI attack iteration will be discarded to ensure that the update of the consistency variables will not be affected by false data. j The consistency variable λ j (k+1), the update process at the kth iteration is expressed as

[0153]

[0154] Combined with the smart grid sensing, transmission and control integrated control architecture proposed in the present invention, and the encryption interaction mode based on hierarchical collaborative DQN and the quantum key distribution rate joint optimization method, the proposed power communication coupled smart grid sensing, transmission and control integrated method process is as follows: Figure 2 As shown, the specific steps are as follows:

[0155] Step 1: Use the Kalman filter method to construct a consistent variable estimate time series X j′,j (k) and the measurement value time series Y j′,j (k);

[0156] Step 2: Calculate X according to formula (7) j′,j (k) and Y j′,j (k) relevance;

[0157] Step 3: Perform FDI attack perception according to formula (8);

[0158] Step 4: Make an encryption interaction mode decision according to formula (9);

[0159] Step 5: Formulate the leader node key distribution rate strategy according to formula (10);

[0160] Step 6: Formulate the follower node key distribution rate strategy according to formula (11);

[0161] Step 7: Calculate the loss function according to formula (12), feed back the evaluation result of the loss function to each node, and update its Q value and quantum key distribution rate strategy;

[0162] Step 8: Calculate the consistency variable importance weight based on encryption communication delay and sequence correlation according to formula (13);

[0163] Step 9: Calculate the normalized importance weight according to formula (14);

[0164] Step 10: According to the FDI attack detection results, discard the attacked consistency variables;

[0165] Step 11: Update the consistency variable according to formula (15);

[0166] Step 12: Repeat steps 1 to 11 to perform integrated optimization of power communication coupled smart grid sensing, transmission and control until the entire optimization cycle is completed, and output the optimal output power P according to formula (3) j .

[0167] The present invention proposes a power communication coupled smart grid sensing, transmission and control integrated device such as Figure 3 As shown, the power communication coupled smart grid sensing, transmission and control integrated device includes an FDI attack perception unit based on sequential Kalman filtering, an encryption interaction method and quantum key distribution rate joint optimization unit based on hierarchical collaborative DQN, and an electrical consistency control optimization unit considering communication delay and attack perception.

[0168] First, the FDI attack perception unit based on time series Kalman filtering includes an estimated value time series calculation module, a measured value time series calculation module, a correlation calculation module, and an FDI attack perception module. The specific functions and functional interactions of each module are shown below.

[0169] Estimated value time series calculation module: The Kalman filtering method is used to calculate the estimated value time series of the consistency variable and the results are transmitted to the correlation calculation module.

[0170] Measurement value time series calculation module: uses the Kalman filtering method to calculate the consistency variable measurement value time series and transmits the result to the correlation calculation module.

[0171] Correlation calculation module: uses cosine similarity to measure the correlation between the optimal estimation value sequence and the measurement value sequence, and transmits the result to the FDI attack perception module.

[0172] FDI attack perception module: FDI attack perception is performed by comparing the correlation between two sequences and comparing them with the threshold, and the results are transmitted to the consistency variable discarding module.

[0173] Secondly, the encryption interaction mode and quantum key distribution rate joint optimization unit based on hierarchical collaborative DQN includes DQN network module, encryption interaction mode decision module, leader node key distribution rate decision module, follower node key distribution rate decision module and loss function calculation module. The specific role and functional interaction of each module are as follows.

[0174] DQN network module: Based on the loss function calculation results, perform agent learning, update the Q value and quantum key distribution rate strategy. Based on the current state information, calculate the state-action evaluation function and transmit the result to the encryption interaction mode decision module, the leader node key distribution rate decision module and the follower node key distribution rate decision module.

[0175] Encryption interaction mode decision module: Based on the DQN network module, it calculates and compares the state-action evaluation function size, makes node encryption interaction mode decisions, and transmits the results to the loss function calculation module and the electrical consistency control optimization unit that considers communication delay and attack perception.

[0176] Leader node key distribution rate decision module: Based on the DQN network module, the leader node key distribution rate is formulated according to the key distribution value function in the Q network, and the result is transmitted to the follower node key distribution rate decision module, the loss function calculation module and the electrical consistency control optimization unit considering communication delay and attack perception.

[0177] Follower node key distribution rate decision module: Based on the DQN network module, the follower node key distribution rate is determined according to the distribution rate of the leader node and its own network status, and the result is transmitted to the loss function calculation module and the electrical consistency control optimization unit considering communication delay and attack perception.

[0178] Loss function calculation module: The loss function is calculated based on the number of encryption failures caused by key resource waste and insufficient quantum keys. After each iteration, the evaluation result of the loss function is fed back to the DQN network module to update its Q value and quantum key distribution rate strategy. Through continuous iteration and optimization, the quantum key distribution rate strategy can be dynamically adjusted.

[0179] Finally, the electrical consistency control optimization unit considering communication delay and attack perception includes an importance weight calculation module, a consistency variable discarding module, a consistency variable updating module based on importance weight and attack perception, and an output power calculation module. The specific functions and functional interactions of each module are shown below.

[0180] Importance weight calculation module: Based on the encryption communication delay and sequence correlation, calculate the importance weight of the consistency variable, normalize the importance weight, and transmit the result to the consistency variable update module based on the importance weight and attack perception.

[0181] Consistency variable discarding module: Based on the FDI attack perception results, the consistency variables detected in the FDI attack are discarded to ensure that the update of the consistency variables will not be affected by false data, and the results are transmitted to the consistency variable updating module based on importance weight and attack perception.

[0182] Consistency variable update module based on importance weight and attack perception: According to the consistency variable importance weight and FDI attack-aware consistency variable discarding results, the consistency variables of each node are updated and the results are transmitted to the output power calculation module.

[0183] Output power calculation module: Calculates the output power of each node based on the update results of the consistency variables.

[0184] The above device is used to carry out integrated optimization of power communication coupled smart grid sensing, transmission and control. The specific steps are as follows.

[0185] Step 0: Initialize each module.

[0186] Step 1: The estimated value time series calculation module and the measured value time series calculation module use the Kalman filtering method to construct the consistent variable estimated value time series X j′,j (k) and the measurement value time series Y j′,j (k);

[0187] Step 2: Correlation calculation module, calculate X j′,j (k) and Y j′,j (k) relevance;

[0188] Step 3: FDI attack perception module, performs FDI attack perception based on the formula correlation calculation results;

[0189] Step 4: The encryption interaction mode decision module calculates and compares the state-action evaluation function size to make a node encryption interaction mode decision;

[0190] Step 5: The leader node key distribution rate decision module determines the leader node key distribution rate according to the key distribution value function in the Q network;

[0191] Step 6: The follower node key distribution rate decision module determines the follower node key distribution rate according to the leader node's distribution rate and its own network status;

[0192] Step 7: The loss function calculation module calculates the loss function according to the number of encryption failures caused by key resource waste and insufficient quantum key. After each iteration, the evaluation result of the loss function is fed back to the DQN network module.

[0193] Step 8: The DQN network module updates its Q value and quantum key distribution rate strategy based on the loss function. Through continuous iteration and optimization, the quantum key distribution rate strategy can be dynamically adjusted;

[0194] Step 9: The importance weight calculation module calculates the importance weight of the consistency variable based on the encryption communication delay and sequence correlation, and normalizes the importance weight;

[0195] Step 10: Calculate the normalized importance weight according to formula (14);

[0196] Step 11: According to the FDI attack detection results, discard the attacked consistency variables;

[0197] Step 12: The consistency variable discarding module discards the consistency variables detected with FDI attacks according to the FDI attack perception results, ensuring that the update of consistency variables will not be affected by false data;

[0198] Step 13: Based on the importance weight and attack-aware consistency variable update module, the consistency variables of each node are iteratively updated according to the consistency variable importance weight and the FDI attack-aware consistency variable discarding result;

[0199] Step 13: The output power calculation module calculates the output power of each node according to the update result of the consistency variable;

[0200] Step 14: Repeat steps 1 to 13 to optimize the integration of power communication coupling and smart grid sensing, transmission and control until the entire optimization cycle is completed and the optimal output power P is output. j .

[0201] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the technical principles of the present invention, and these improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for integrating power communication coupling with smart grid sensing, transmission and control, characterized in that: The specific steps are: Firstly, an electrical control model based on a distributed consensus algorithm is proposed to reduce the dependence of control on the central node. Secondly, in response to the FDI attack in the system, a communication control model integrating quantum encryption and classical encryption is proposed to dynamically adjust the encryption method and key distribution rate to ensure data security while reducing the transmission delay of the consistent variable of quantum encryption; Next, a communication-electricity collaborative control model under FDI attack is proposed. Based on FDI attack perception, the encryption interaction mode and key distribution rate of the communication domain and the importance weight of the consistency variables of the power domain are dynamically adjusted to prevent the injected false data from spreading through the communication network and achieve fast and accurate convergence of consistency variables. Based on the above three models, the time series Kalman filtering method is used to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of consistency variables and the measurement value sequence, ensuring the authenticity and reliability of consistency variable data; In terms of communication control, when an FDI attack is detected, the system switches the encryption interaction mode to quantum encryption based on the state-action evaluation function in the Q network, and selects and dynamically adjusts the respective quantum key distribution rates through the collaborative work of the leader node and the follower node; Finally, in terms of electrical control, the importance weights of consistency variables are constructed based on the serial correlation of encrypted communication delay and attack perception, and normalized, the consistency variables are updated, and the integrated optimization of power communication coupled smart grid sensing, transmission and control is carried out until the entire optimization cycle is completed and the optimal output power is output.

2. The method for integrating sensing, transmission and control of power communication coupled smart grid according to claim 1, characterized in that: The electrical control model based on the distributed consistency algorithm is specifically: Define the number of fusion terminals as J, and its set is At the jth fusion terminal s j The output control cost function F of the distributed generation in the controlled area j is j (P j )for: In the formula, α j , γ j , β j They represent the quadratic coefficient, linear coefficient, and constant coefficient of the distributed generation output cost function in the area j, respectively. j is the output power of distributed power sources in area j; Based on the Lagrange multiplier method and KKT conditions, the consistency variable λ is introduced j , represents the Lagrange multiplier of the control area controlled by the j-th fusion terminal sj; The definition of consistency control process can be divided into K iterations, and the consistency variable λ of the jth fusion terminal j (k+1), the update process at the kth iteration is expressed as: In the formula, ξ j (k) represents the number of λ in the kth iteration j (k) is the update step size; τ j′,j (k) is the number of s in the kth iteration j′ to j The communication delay of j′,j (k) is the kth iteration s j′ to j The importance weight of the consistency variables; λ j′ (k-τ j′,j (k)) represents the consideration of communication delay τ j′,j (k) Next, s j From j′ Received consistency variable; j-th fusion terminal s j The consistency variable λ of the controlled area j The convergence value of the optimal distributed power output power P corresponding to the area j j , expressed as 3. The method for integrating sensing, transmission and control of power communication coupled smart grid according to claim 1, characterized in that: The communication control model integrating quantum encryption and classical encryption is specifically as follows: definition is the encrypted interaction mode between fusion terminals; j,j′ (k) = 1 means s j With s j′ Quantum encryption is used to interact with each other; j,j′ (k) = 0 means that the classic encryption method is used for interaction; definition and are the quantum node set and the classical node set respectively; define τ j,j′ (k) is the number of s in the kth iteration j With s j′ The encrypted communication delay between is expressed as: In the formula, and Respectively represent the k-th iteration s j With s j′ The communication delay between classical encryption and quantum encryption.

4. The method for integrating sensing, transmission and control of power communication coupled smart grid according to claim 1, characterized in that: The communication-electrical coordinated control model under the FDI attack is specifically as follows: During the consistency control process, the FDI attacker intercepts s j With s j′ The classical channel data between them is injected with false data to destroy the convergence of the consistency variables and reduce the control performance; define is the number of times the attacker is in s in the kth iteration j With s j′ The false data injected during the tampering is expressed as: In the formula, Θ j′,j (k) represents the kth iteration of the attacker in s j With s j′ Indicator variable for inter-FDI attacks, Θ j′,j (k) = 1 means s j With s j′ The communication data between them is attacked by FDI, otherwise Θ j′,j (k) = 0; Communication-electrical cooperative control model under FDI attack, consistency variable λ of the jth fusion terminal j (k+1), the update process at the kth iteration is expressed as:

5. The method for integrating sensing, transmission and control of power communication coupled smart grid according to claim 1, characterized in that: The specific method of using the time series Kalman filtering method to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of the consistency variable and the measurement value sequence is as follows: Construct s in the kth iteration j With s j′ The optimal estimate of the consistency variable between time series X j′,j (k) = [X j′,j (kk * ),…,X j′,j (k-1),X j′,j (k)] and the measurement value time series Y j′,j (k) = [Y j′,j (kk * ),…,Y j′,j (k-1),Y j′,j (k)]; FDI attack perception is performed by comparing the correlation of the two sequences and comparing them with the threshold; the cosine similarity is used to measure the correlation between the optimal estimation value sequence and the measurement value sequence; in the kth iteration, s j With s j′ The correlation between j′,j (k) and FDI attack perception are expressed as: In the formula, ||.|| represents the Euclidean norm; X j′,j (k) = [X j′,j (kk * ),…,X j′,j (k-1),X j′,j (k)] represents the optimal estimate sequence, whose elements include the (kk * ) iterations to the optimal estimate of the kth iteration; Y j′,j (k) = [Y j′,j (kk * ),…,Y j′,j (k-1),Y j′,j (k)] represents the measurement value sequence; ΔF represents the FDI attack perception threshold; Indicates that s in the kth iteration j With s j′ The indicator variable of FDI attack perception between Indicates the perception of j With s j′ There is an FDI attack; conversely, 6. The method for integrating sensing, transmission and control of power communication coupled smart grid according to claim 1, characterized in that: The system switches the encryption interaction mode to quantum encryption mode based on the state-action evaluation function in the Q network, and through the collaborative work of the leader node and the follower node, the specific method of selecting and dynamically adjusting the respective quantum key distribution rates is as follows: The fusion terminal that directly interacts with the distribution network control center is defined as the leader node, and the set is represented as Other terminals that cannot directly interact with each other are follower nodes, and the set is represented as To ensure consistent convergence, the leader node and the quantum nodes with which it has communication connections use quantum encryption interaction; Follow Nodes j Based on the Q network, construct the state-action evaluation function Q(Φ j (k),s j′ ,y j′,j (k)); Q(Φ j (k),s j′ ,y j′,j (k)=1) means that in state Φ j (k) Next j With s j′ The value of quantum encryption interaction between j (k),s j′ ,y j′,j (k) = 0) represents the value of using the classic encryption interaction method; therefore, the encryption interaction method decision is expressed as: y j′,j (k)=Ξ{Q(Φ j (k),s j′ ,1)>Q(Φ j (k),s j′ ,0)} (9) Based on the hierarchical collaborative DQN algorithm, the quantum key distribution rate is dynamically adjusted according to the interaction requirements, the node's own distribution rate and upper limit, and the key inventory. First, the leader node selects the key distribution rate based on the key distribution value function in the Q network. Second, the follower node makes further adjustments based on the leader node's adjusted distribution rate and its own network status to ensure that the distribution rate is consistent with the neighboring nodes in the interaction. Finally, multiple iterations and feedback optimization are used to gradually approach the optimal strategy. The leader node s of the kth iteration j With follower nodes s j′ The initial key distribution rate a between j,j′ (k) is expressed as: In the formula, Distribute value functions for keys; Represents the leader node s j To all adjacent follower nodes s j′ The sum of the values ​​corresponding to the key distribution; Specify the node s to follow in the kth iteration j′ The key distribution rate It is expressed as: In the formula, for The distribution rate and upper limit of Representation and There is an interactive leader node; the first part in brackets is the remaining available distribution rate; in the numerator, express In Status Next, select j′ The Q value corresponding to the key distribution; the denominator is The sum of the Q values ​​of other follower nodes that interact; Construct a loss function to evaluate the joint optimization performance of encryption interaction mode and key distribution rate; define the kth iteration leader node s j The loss variable is R j (k), to evaluate the number of encryption failures caused by key resource waste and insufficient quantum keys, and to construct the loss function L j for: Where K is the number of iterations at convergence, γ j is a discount factor used to evaluate the impact of future rewards on current decisions; Indicates that in the k+1th iteration, the leader node s j In the state Φ j Select the best follower node under (k+1) The maximum Q value corresponding to key distribution; After each iteration, the evaluation results of the loss function are fed back to each node to update its Q value and quantum key distribution rate strategy. Through continuous iteration and optimization, dynamic adjustment of the quantum key distribution rate strategy is achieved.

7. The method for integrating sensing, transmission and control of power communication coupled smart grid according to claim 1, characterized in that: Based on the serial correlation of encrypted communication delay and attack perception, the importance weight of consistency variables is constructed, and normalized, the consistency variables are updated, and the integrated optimization of power communication coupling smart grid sensing, transmission and control is performed until the end of the entire optimization cycle. The specific method of outputting the optimal output power is: An electrical consistency control optimization method considering communication delay and attack perception is proposed. After detecting an FDI attack, the consistency variable importance weight based on the encryption communication delay and sequence correlation is constructed according to the delay of the received message and the encryption interaction mode. The encryption interaction mode switching and elastic step size are used to prevent false signals from spreading to the global network through the communication bridge and causing cascading oscillations. Considering the encryption communication delay and attack perception, construct the kth iteration s j With s j′ The importance weight of the consistent variables between is expressed as: Where Δ(k) is the elastic step size of the kth iteration, which gradually approaches 0 as the number of iterations increases, ensuring that the adjustment amplitude is large in the initial stage of the algorithm and becomes smaller in the later stage of the iteration to ensure consistent convergence; φ τ With φ F are adjustment factors, which are used to adjust the impact of encryption communication delay and sequence correlation on importance weight; The weights are normalized and expressed as: When j′ to j When there is an FDI attack, the consistency variables detected during the FDI attack iteration will be discarded to ensure that the update of the consistency variables will not be affected by false data. j The consistency variable λ j (k+1), the update process at the kth iteration is expressed as:

8. An integrated device for sensing, transmission and control of power communication coupled smart grid, characterized in that: include: Modeling unit: used to first propose an electrical control model based on a distributed consensus algorithm to reduce the control's dependence on the central node; Secondly, in response to the FDI attack in the system, a communication control model integrating quantum encryption and classical encryption is proposed, which dynamically adjusts the encryption method and key distribution rate to ensure data security while reducing the transmission delay of the consistency variables of quantum encryption; then, a communication-electricity collaborative control model under FDI attack is proposed, which dynamically adjusts the encryption interaction method and key distribution rate of the communication domain and the importance weight of the consistency variables of the power domain based on FDI attack perception to avoid the spread of injected false data through the communication network and achieve rapid and accurate convergence of consistency variables; FDI attack perception unit based on time series Kalman filtering: Based on the above three models, the time series Kalman filtering method is used to achieve accurate perception of FDI attacks by comparing the correlation between the optimal estimation sequence of the consistency variable and the measurement value sequence, ensuring the authenticity and reliability of the consistency variable data; Joint optimization unit of encryption interaction mode and quantum key distribution rate based on hierarchical collaborative DQN: used in communication control. When an FDI attack is detected, the system switches the encryption interaction mode to quantum encryption mode based on the state-action evaluation function in the Q network, and selects and dynamically adjusts the respective quantum key distribution rates through the collaborative work of the leader node and the follower node; Electrical consistency control optimization unit considering communication delay and attack perception: In terms of electrical control, it is used to construct consistency variable importance weights based on the serial correlation of encrypted communication delay and attack perception, and perform normalization, update consistency variables, and perform integrated optimization of power communication coupled smart grid sensing, transmission and control until the end of the entire optimization cycle and output the optimal output power.

9. The power communication coupled smart grid sensing, transmission and control integrated device according to claim 8, characterized in that: The FDI attack perception unit based on temporal Kalman filtering includes: Estimated value time series calculation module: uses the Kalman filtering method to calculate the estimated value time series of the consistency variable and transmits the result to the correlation calculation module; Measurement value time series calculation module: uses Kalman filtering method to calculate the consistency variable measurement value time series, and transmits the result to the correlation calculation module; Correlation calculation module: uses cosine similarity to measure the correlation between the optimal estimation value sequence and the measurement value sequence, and transmits the result to the FDI attack perception module; FDI attack perception module: FDI attack perception is performed by comparing the correlation between two sequences and comparing them with the threshold, and the results are transmitted to the consistency variable discarding module.

10. The power communication coupled smart grid sensing, transmission and control integrated device according to claim 8, characterized in that: The encryption interaction mode and quantum key distribution rate joint optimization unit based on hierarchical collaborative DQN includes: DQN network module: Based on the loss function calculation results, perform intelligent agent learning, update the Q value and quantum key distribution rate strategy. Based on the current state information, calculate the state-action evaluation function and transmit the result to the encryption interaction mode decision module, the leader node key distribution rate decision module and the follower node key distribution rate decision module; Encryption interaction mode decision module: Based on the DQN network module, it calculates and compares the state-action evaluation function size, makes node encryption interaction mode decisions, and transmits the results to the loss function calculation module and the electrical consistency control optimization unit that considers communication delay and attack perception; Leader node key distribution rate decision module: Based on the DQN network module, the leader node key distribution rate is determined according to the key distribution value function in the Q network, and the result is transmitted to the follower node key distribution rate decision module, the loss function calculation module and the electrical consistency control optimization unit considering communication delay and attack perception; Follower node key distribution rate decision module: Based on the DQN network module, it determines the follower node key distribution rate according to the leader node's distribution rate and its own network status, and transmits the result to the loss function calculation module and the electrical consistency control optimization unit considering communication delay and attack perception; Loss function calculation module: The loss function is calculated based on the number of encryption failures caused by key resource waste and insufficient quantum keys. After each iteration, the evaluation result of the loss function is fed back to the DQN network module to update its Q value and quantum key distribution rate strategy. Through continuous iteration and optimization, dynamic adjustment of the quantum key distribution rate strategy is achieved.

11. The power communication coupled smart grid sensing, transmission and control integrated device according to claim 8, characterized in that: The electrical consistency control optimization unit considering communication delay and attack perception includes: Importance weight calculation module: calculates the importance weight of consistency variables based on encryption communication delay and sequence correlation, normalizes the importance weight, and transmits the result to the consistency variable update module based on importance weight and attack perception; Consistency variable discarding module: based on the FDI attack perception results, the consistency variables detected by FDI attack are discarded to ensure that the update of consistency variables will not be affected by false data, and the results are transmitted to the consistency variable updating module based on importance weight and attack perception; Consistency variable update module based on importance weight and attack perception: updates the consistency variables of each node according to the consistency variable importance weight and the consistency variable discarding result of FDI attack perception, and transmits the result to the output power calculation module; Output power calculation module: Calculates the output power of each node based on the update results of the consistency variables.

12. An integrated system of power communication coupled smart grid sensing, transmission and control, characterized in that: It includes electrical control layer, classical channel transmission layer, quantum key distribution layer and communication control layer; The electrical control layer includes multiple distribution substations, each of which is optimized by a fusion terminal for electrical control decisions, and controls the output of distributed power sources to meet the energy supply and demand balance; The classical channel transmission layer is composed of 5G channels and is used to transmit data encrypted by quantum keys or classical keys; The quantum key distribution layer is composed of quantum channels and is used to distribute quantum keys.

Images