Encryption transmission and self-healing control cooperation system facing virtual power plant terminal

Abstract

The invention discloses a virtual power plant terminal-oriented encryption transmission and self-healing control cooperative system, which comprises a dynamic encryption module, a hierarchical encryption system is constructed on the basis of a quantum chaos sequence and a national secret SM9 algorithm, a dynamic session key is generated in real time through an edge node, and terminal power data is subjected to hierarchical encryption; the self-healing control module integrates a block chain distributed account book and a deep reinforcement learning model, constructs a health degree portrait of the terminal equipment in real time, monitors a link state through a topology sensing algorithm, predicts a fault propagation path, and triggers communication link reconstruction and computing resource elastic scheduling for self-healing; the collaborative optimization module constructs an optimization objective function, balances encryption strength and self-healing efficiency, adopts a quantum topology photonic crystal optimization algorithm to determine a weight coefficient in the optimization objective function, and realizes global optimal matching of an encryption-self-healing strategy; and the edge computing node performs terminal-edge cloud cooperative training according to the global optimal matching strategy of the encryption-self-healing strategy.

Description

Technical Field

[0001] The present invention relates to the cross - field of power system information security and edge - computing intelligent control, and particularly relates to a collaborative system for encrypted transmission and self - healing control for virtual power plant terminals, which uses cryptography to encrypt the operation of virtual power plants, evaluate their health conditions and perform self - healing. Background Art

[0002] Virtual power plant terminal devices need to transmit sensitive control instructions and operation data frequently in power dispatching, demand response and distributed energy management, facing risks such as network attacks, channel interference and equipment anomalies. However, data encryption and fault self - healing mechanisms mostly operate independently, which is likely to cause high latency in transmission, resulting in a lag in self - healing response. At the same time, static key management and centralized control architectures are difficult to adapt to dynamic network topology changes, posing potential risks of data leakage and system paralysis. The present invention realizes closed - loop control of secure transmission and real - time recovery through an encryption - self - healing joint optimization model, improving the security and stability of virtual power plant terminals in complex environments. Summary of the Invention

[0003] Object of the Invention: Aiming at the multiple security threats and equipment fault recovery requirements of virtual power plant terminals in an open network environment, the present invention proposes a collaborative system for encrypted transmission and self - healing control for virtual power plant terminals, which ensures data confidentiality and integrity through dynamic cryptography strategies and edge - side intelligent decision - making mechanisms, and realizes rapid self - healing of abnormal equipment states, ensuring the safe and stable operation of the power system.

[0004] Technical Solution: The present invention discloses a collaborative system for encrypted transmission and self - healing control for virtual power plant terminals, which is characterized by including a dynamic encryption module, a self - healing control module, a collaborative optimization module and an edge - computing node;

[0005] The dynamic encryption module constructs a hierarchical encryption system based on quantum chaotic sequences and the national standard SM9 algorithm, generates dynamic session keys in real - time through edge nodes, and encrypts terminal power data in grades. The hierarchical encryption system includes: a quantum chaotic sequence generator, a key derivation function, and multi - level encryption channels. The quantum chaotic sequence generator generates an initial entropy source according to the terminal device fingerprint and environmental noise, generates a chaotic sequence through non - linear feedback modulation. The key derivation function performs an exclusive - OR operation on the bilinear pair mapping in the SM9 algorithm and the chaotic sequence to generate a dynamic key seed, and expands the seed into multi - level session keys through the HKDF standard. The multi - level encryption channels encrypt data according to different encryption strategies based on data sensitivity levels;

[0006] The self-healing control module integrates the blockchain distributed ledger and the deep reinforcement learning DRL model, constructs the health profile of the terminal device in real time, monitors the link status through the topology perception algorithm, predicts the fault propagation path in combination with the Lyapunov stability criterion, and triggers the communication link reconstruction and the elastic scheduling of computing resources for self-healing;

[0007] The collaborative optimization module constructs a multi-objective optimization function to balance the encryption strength and the self-healing efficiency, and uses the quantum topology photonic crystal optimization algorithm to determine the numerical values of the multi-objective optimization function indicators, realizing the global optimal matching of the encryption-self-healing strategy; the optimization objective function includes four indicators: encryption delay, data integrity loss, self-healing response time, and resource utilization;

[0008] The edge computing node deploys the federated learning framework, and according to the global optimal matching strategy of the encryption-self-healing strategy, conducts terminal-edge cloud collaborative training, dynamically updates the encryption policy library and the self-healing rule library, supports local decision-making in the offline mode and global optimization in the online mode; the edge node builds a lightweight consensus mechanism inside, and through local model gradient aggregation and global parameter synchronization, ensures the consistency of the encryption policy and the self-healing rule.

[0009] Furthermore, the data sensitivity level is divided into three levels of data, and the differential encryption strategies are as follows: the first-level data uses the composite encryption of SM9 identity encryption and chaotic sequence stream cipher; the second-level data uses the dynamic message authentication code generated by adding a chaotic sequence after SM9 encryption; the third-level data only uses the basic SM9 encryption, and the session key update period is synchronized with the power data acquisition frequency.

[0010] Furthermore, the blockchain distributed ledger and the deep reinforcement learning DRL model are integrated to construct the health profile of the terminal device in real time, specifically as follows:

[0011] 1) The blockchain distributed ledger records the operating state parameters such as the voltage volatility δ v , temperature drift ΔT, communication error rate ε of the terminal device and the encryption link integrity score, and constructs a tamper-proof data chain through the Merkle tree;

[0012] 2) The health profile is constructed based on the statistical physics model of multi-dimensional state parameters, which is specifically divided into: feature extraction, state dimensionality reduction, health calculation, and dynamic optimization;

[0013] Feature extraction: Extract the time-domain feature mean and variance of the device historical data, the frequency-domain feature wavelet packet energy entropy E T and the non-linear feature Lyapunov exponent λ, forming an 8-dimensional feature vector F(t)∈ 8 ;

[0014] State dimension reduction: Unsupervised dimension reduction of F(t) is performed through a deep autoencoder to generate a 3D latent health state Z(t) ∈ 3 ;

[0015] Health degree calculation: Define the health index as: where μ z and ∑ z are the mean and covariance matrix of the latent state under normal working conditions, Dynamically optimize the DRL model driven by the dual Q network, and dynamically optimize the health assessment strategy with the reward function. Define the reward function as: where is the fault label function.

[0016] Furthermore, monitor the link state through the topology awareness algorithm, predict the fault propagation path in combination with the Lyapunov stability criterion, and trigger communication link reconstruction and elastic scheduling of computing resources for self-healing, as follows:

[0017] 1) The topology awareness algorithm monitors the power communication link and the physical power link in real time and constructs a cross-layer link state matrix. The power communication link is the blockchain data transmission channel, and the parameters include the bit error rate ∈ j (t), the dynamic error correction code recovery efficiency r j (t), the number of active data blocks N j (t), the time delay τ j (t), and the bandwidth B j (t); the physical power link is the power equipment connection line, and the parameters include the line power flow P j (t), the voltage volatility δ v (t), the temperature drift ΔT(t), and the on-off state C j (t) between nodes. The cross-layer link state matrix Q(t) ∈ m×5 is:

[0018]

[0019] where m is the total number of links, and j = 1, 2, … m;

[0020] 2) The Lyapunov stability predicts the fault propagation path of the power physical line and the fault propagation path of the communication link:

[0021] According to the power flow equation of the power line calculate the fault propagation path length L path (t). When is detected, trigger the communication link reconstruction based on SDN, and switch to the redundant path and calculate the elastic resource scheduling within T reconfig < 50 ms, and perform according to C usedC(t) = min(1.2C base , C total ) dynamically allocates edge node resources. If the physical link C j (t) = 0, then synchronously increase the encryption strength coefficient k i (t) ≥ 0.8 and start data retransmission, where T reconfig is the communication link reconstruction time, C base is the benchmark computing resource requirement, C total is the total available computing resource.

[0022] Furthermore, the multi-objective optimization function F of the collaborative optimization module is specifically as follows:

[0023]

[0024] f = [f1, f2, f3, f4] T

[0025]

[0026]

[0027] where f = [f1, f2, f3, f4] T is the factor vector, w = [w1, w2, w3, w4] T is the weight vector, represents the Hadamard product, and the weights are dynamically allocated through the quantum topological photonic crystal optimization algorithm. f1 is the encryption delay efficiency factor, f2 is the data integrity factor, f3 is the self-healing response efficiency factor, f4 is the resource utilization efficiency factor, w1 is the weight coefficient of encryption delay, w2 is the weight coefficient of data integrity loss, w3 is the weight coefficient of self-healing response time, w4 is the weight coefficient of resource utilization rate, T enc is the encryption delay, L int is the data integrity loss, T heal is the self-healing response time, U res is the resource utilization rate, is the encryption delay reference value, is the data integrity reference value, U base is the resource utilization rate reference value, D i is the size of the i-th level data packet, v max is the maximum encryption rate, ∈ max is the maximum allowable bit error rate, N max is the maximum number of active data chunks, C total is the total computing resource, p is the total number of resource types, D i (t) is the instantaneous size of the i-th level data packet at time t, k i(t) is the dynamic encryption strength coefficient, v(t) is the real-time encryption rate of edge nodes, Δt is the time window of the encryption session, ∈ j (t) is the instantaneous bit error rate of the j-th link at time t, r j (t) is the dynamic error correction code recovery efficiency of the j-th link at time t, N j (t) is the number of active data blocks of the j-th link at time t, is the instantaneous change rate of the fault propagation path length, M(t) is the real-time number of available redundant links, α and β are system dynamic adjustment coefficients, is the instantaneous utilization rate of the q-th type of resource, C q,total is the total amount of the q-th type of resource.

[0028] Further, the quantum topology photonic crystal optimization algorithm QTPC is adopted to determine the numerical values of the multi-objective optimization function indicators, and the global optimal matching of the encryption-self-healing strategy is realized. The process is as follows:

[0029] (1) In the quantum topology initialization stage, the range threshold and the maximum number of iterations are set. Based on the quantum chaotic sequence generator of the dynamic encryption module, a topologically protected quantum state is constructed in the periodic structure of the photonic crystal. Through the valley Hall effect, the hierarchical characteristics of power data are mapped to the high-symmetry path in the momentum space, and an initial quantum state carrying the encryption key seed is generated. The formula is:

[0030]

[0031] Among them, ψ i (0) is the wave function of the i-th quantum topology state at the initial moment, representing a candidate solution to the optimization problem, is the quantum topology photonic crystal encoding operator, which maps the power data characteristics to the energy band structure of the photonic crystal, k m is the m-th sampling point in the momentum space, corresponding to the hierarchical characteristics of the power data at the virtual power plant terminal, r is the spatial coordinate position of the photonic crystal, χ v,K (r) is the wave function of the topological boundary state with Chern number v = 1, K is the high-symmetry point in the momentum space, providing topological protection characteristics, M is the total number of sampling points in the momentum space, which is consistent with the number of power data hierarchies. The dynamic parameter S output by the quantum chaotic sequence generator SM9 (t) modulates the phase φ of the defect mode of the photonic crystal m , so that the key generation process forms quantum entanglement with the topological energy band structure. The lattice constant a and the encryption strength coefficient k i (t) satisfy Δ gap = a·k i (t), realizing the adaptive adjustment of the encryption strength at the physical layer;

[0032] (2) Anomalous quantum Hall evolution stage, combined with the real-time link state data N j (t) and the bit error rate ∈ j (t), construct a dynamic artificial gauge field to drive the Berry curvature evolution, and achieve dissipationless optimization search through the quantum anomalous Hall effect. The specific dynamic equation is:

[0033]

[0034] Among them, e is the elementary charge quantity, is the reduced Planck constant, ε(N j (t)) is the virtual electric field strength, dynamically adjusted by the number of active data blocks N j (t) of the j-th link, Ω(k) is the Berry curvature tensor, is the momentum space gradient operator, driving the quantum state to evolve towards the low energy direction, β(∈ j (t)) is the adaptive diffusion coefficient, negatively correlated with the instantaneous bit error rate ∈ j (t) of the j-th link, is the momentum space Laplace operator, controlling the diffusion range of the quantum state. When N j (t) decreases, the electric field is enhanced to expand the search range. When ∈ j (t) increases, the diffusion coefficient β is suppressed to prevent local convergence, forming a closed-loop balance between encryption reliability and self-healing response speed;

[0035] (3) Photo-acoustic coupling weight assignment stage. Through the strong coupling between surface plasmons and phonon modes, map the multi-objective function of the collaborative optimization module to the quantum degenerate state. The Rabi oscillation frequency is jointly determined by the power data grading parameter D i (t) and the dynamic encryption intensity coefficient k i (t), and realize the dynamic physical assignment of the weight coefficient. The formula is:

[0036]

[0037] Among them, j = 1 corresponds to the encryption delay T enc , j = 2 corresponds to the data integrity loss L int , j = 3 corresponds to the self-healing response time T heal , j = 4 corresponds to the resource utilization rate U res ; The photo-acoustic coupling intensity g is positively correlated with the computing resource C q,used of the edge node, ensuring that the weight assignment process adapts to changes in resource load. w j is the weight coefficient of the j-th optimization objective, D i (t) is the instantaneous size of the i-th level data packet at time t, k i(t) The dynamic encryption strength coefficient of the i-th level data, with a value range of [0.1, 1.0], is generated by the SM9 algorithm;

[0038] (4) In the dynamic topology reconstruction stage, based on the gradient matrix W output by the edge computing node federated learning framework FL , the dielectric constant distribution of the photonic crystal is reconstructed in real time to make the optimization process adapt to the topology changes of the power network. The formula is:

[0039]

[0040] where ∈0 is the basic dielectric constant, ReLU is the rectified linear unit activation function, is the partial derivative of the multi-objective optimization function F with respect to the weight w j , reflecting the sensitivity of each objective. The federated gradient W synchronized by the lightweight consensus mechanism FL After being activated by the rectified linear unit, dynamically adjusts the dielectric constant ∈(r, t) to form the co-evolution of the encryption policy library and the self-healing rule library. The topological protection energy gap width Δ gap and the resource utilization rate Ures show a negative correlation characteristic, realizing the quantization trade-off between security and efficiency;

[0041] (5) In the quantum error correction output stage, the surface code quantum error correction protocol is used to perform anti-interference coding on the optimization solution set to support the reliability of local decision-making in the offline mode;

[0042] (6) Termination condition determination: The optimization iteration is terminated when any of the following conditions is met:

[0043] 1. Hard constraints of the power system: T enc <10ms and T heal <50ms;

[0044] 2. Quantum fidelity threshold: Tr(ρ out ρ ideal )>0.9999;

[0045] 3. Resource protection mechanism: U res >95% for 1s continuously.

[0046] Furthermore, optimize the algorithm optimization process, and use the quantum tunneling effect to break through the traditional solution space limit. The formula is:

[0047]

[0048] where the tunneling coefficient k = 0.1 / ms is synchronized with the key update event to ensure the time consistency of the encryption parameters and the resource scheduling strategy, is the original solution without boundary processing, L j , U jare the lower and upper bounds of the j-th optimization variable, k is the tunneling coefficient, which controls the convergence speed when the solution vector exceeds the boundary, and T refresh is the SM9 key update timestamp, which is synchronized with the dynamic encryption module to ensure timing consistency.

[0049] Beneficial effects

[0050] 1. The present invention constructs a dynamic hierarchical encryption mechanism by integrating quantum chaotic sequences and national cryptographic algorithms, realizes real-time key update and data hierarchical protection, significantly reduces communication delay while enhancing the anti-quantum attack ability; constructs a device health assessment system by combining blockchain trusted storage and intelligent learning models, and accurately predicts the fault propagation path based on stability criteria, effectively shortening the system self-healing response time and improving the abnormal positioning accuracy.

[0051] 2. The present invention improves the multi-objective optimization algorithm to balance encryption strength and resource efficiency, optimizes the system resource utilization rate on the premise of ensuring data integrity, and supports collaborative decision-making of large-scale terminal nodes. The edge computing framework based on federated learning realizes fast local decision-making in the offline mode and online global optimization through a distributed policy synchronization mechanism, forming a closed-loop control system of secure transmission, intelligent diagnosis and elastic recovery, and comprehensively enhancing the security protection and self-recovery ability of virtual power plant terminals in complex network environments. Description of the drawings

[0052] Figure 1 is the structural framework diagram of the present invention;

[0053] Figure 2 is the flow chart of the quantum topological photonic crystal optimization algorithm;

[0054] Figure 3 is the work flow chart. Detailed implementation manners

[0055] The present invention will be further described below with reference to the drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and cannot be used to limit the protection scope of the present invention.

[0056] The present invention discloses a collaborative system for encrypted transmission and self-healing control for virtual power plant terminals. Refer to Figure 1 , which includes a dynamic encryption module, a self-healing control module, a collaborative optimization module, and edge computing nodes;

[0057] The dynamic encryption module constructs a hierarchical encryption system based on quantum chaotic sequences and the national cryptography SM9 algorithm. It generates dynamic session keys in real time through edge nodes and encrypts terminal power data at different levels. The hierarchical encryption system includes: a quantum chaotic sequence generator, a key derivation function, and multiple-level encryption channels. The quantum chaotic sequence generator generates an initial entropy source based on the terminal device fingerprint and environmental noise, and produces a chaotic sequence through non-linear feedback modulation. The key derivation function performs an exclusive OR operation on the bilinear pair mapping in the SM9 algorithm and the chaotic sequence to generate a dynamic key seed, and expands the seed into multiple-level session keys through the HKDF standard. The multiple-level encryption channels encrypt data according to different encryption strategies based on the data sensitivity level.

[0058] The data sensitivity level is divided into three levels of data, and the differential encryption strategy is as follows: For level 1 data, a composite encryption of SM9 identity encryption and chaotic sequence stream cipher is used; for level 2 data, a dynamic message authentication code generated by appending a chaotic sequence after SM9 encryption is used; for level 3 data, only the basic SM9 encryption is used, and the session key update period is synchronized with the power data collection frequency.

[0059] The self-healing control module integrates a blockchain distributed ledger and a deep reinforcement learning DRL model, constructs a real-time health profile of terminal devices, monitors the link status through a topology awareness algorithm, predicts the fault propagation path in combination with the Lyapunov stability criterion, and triggers communication link reconstruction and elastic scheduling of computing resources for self-healing.

[0060] Integrating a blockchain distributed ledger and a deep reinforcement learning DRL model, a real-time health profile of terminal devices is constructed as follows:

[0061] 1) The blockchain distributed ledger records the operating state parameters of the voltage volatility δ v , temperature drift ΔT, communication error rate ε of the terminal device and the integrity score of the encrypted link, and constructs an anti-tampering data chain through a Merkle tree.

[0062] 2) The health profile is constructed based on a statistical physics model of multi-dimensional state parameters, which is specifically divided into: feature extraction, state dimensionality reduction, health degree calculation, and dynamic optimization.

[0063] Feature extraction: Extract the time-domain feature mean and variance from the historical data of the device, the frequency-domain feature wavelet packet energy entropy E T and the non-linear feature Lyapunov exponent λ, and form an 8-dimensional feature vector F(t) ∈ 8 .

[0064] State dimensionality reduction: Perform unsupervised dimensionality reduction on F(t) through a deep autoencoder to generate a 3-dimensional latent health state Z(t) ∈ 3 .

[0065] Health degree calculation: Define the health index as: where μ z and ∑ z are the mean and covariance matrix of the potential state under normal working conditions,

[0066] The DRL model driven by the dynamic optimization design double Q network dynamically optimizes the health assessment strategy with the reward function. Define the reward function as: where is the fault label function.

[0067] Monitor the link state through the topology-aware algorithm, predict the fault propagation path in combination with the Lyapunov stability criterion, and trigger communication link reconstruction and elastic scheduling of computing resources for self-healing, as follows:

[0068] 1) The topology-aware algorithm monitors the power communication link and the physical power link in real time and constructs a cross-layer link state matrix. The power communication link is the blockchain data transmission channel, and the parameters include the bit error rate ∈ j (t), the dynamic error correction code recovery efficiency r j (t), the number of active data chunks N j (t), the delay τ j (t), and the bandwidth B j (t); the physical power link is the power equipment connection line, and the parameters include the line power flow P j (t), the voltage volatility δ v (t), the temperature drift ΔT(t), and the on-off state C j (t) between nodes. The cross-layer link state matrix Q(t) ∈ m×5 is:

[0069]

[0070] where m is the total number of links, and j = 1, 2,... m.

[0071] 2) The Lyapunov stability predicts the fault propagation path of the power physical line and the fault propagation path of the communication link:

[0072] According to the power flow equation of the power line calculate the fault propagation path length L path (t). When detecting , trigger the communication link reconstruction based on SDN, and switch to the redundant path and calculate the elastic resource scheduling within T reconfig < 50ms, and dynamically allocate the edge node resources according to C used (t) = min(1.2C base , C total ). If the physical link Cj If \(t = 0\), then the encryption strength coefficient \(k\) is increased synchronously. i If \(t)\geq0.8\) and data retransmission is initiated, where \(T\) reconfig is the communication link reconstruction time, \(C\) base is the benchmark computing resource requirement, \(C\) total is the total available computing resource.

[0073] The collaborative optimization module constructs a multi-objective optimization function to balance encryption strength and self-healing efficiency. The quantum topological photonic crystal optimization algorithm is used to determine the numerical values of the multi-objective optimization function indicators, realizing the global optimal matching of the encryption-self-healing strategy; the optimization objective function includes four indicators: encryption delay, data integrity loss, self-healing response time, and resource utilization rate.

[0074] The multi-objective optimization function \(F\) of the collaborative optimization module is specifically as follows:

[0075]

[0076] \(f=[f_1,f_2,f_3,f_4]\) T

[0077]

[0078] where \(f = [f_1,f_2,f_3,f_4]\) T is the factor vector, \(w=[w_1,w_2,w_3,w_4]\) T is the weight vector, represents the Hadamard product, and the weights are dynamically allocated by the quantum topological photonic crystal optimization algorithm. \(f_1\) is the encryption delay efficiency factor, \(f_2\) is the data integrity factor, \(f_3\) is the self-healing response efficiency factor, \(f_4\) is the resource utilization efficiency factor, \(w_1\) is the weight coefficient of the encryption delay, \(w_2\) is the weight coefficient of the data integrity loss, \(w_3\) is the weight coefficient of the self-healing response time, \(w_4\) is the weight coefficient of the resource utilization rate, \(T\) enc is the encryption delay, \(L\) int is the data integrity loss, \(T\) heal is the self-healing response time, \(U\) res is the resource utilization rate, is the encryption delay reference value, is the data integrity reference value, \(U\) base is the resource utilization rate reference value, \(D\) i is the size of the \(i\)-th level data packet, \(v\) max is the maximum encryption rate, \(\epsilon\) max is the maximum allowable bit error rate, \(N\) max is the maximum number of active data chunks, \(C\) total is the total computing resource, \(p\) is the total number of resource types, \(D\) i(t) is the instantaneous size of the i-th level data packet at time t, k i (t) is the dynamic encryption intensity coefficient, v(t) is the real-time encryption rate of the edge node, Δt is the time window of the encryption session, ∈ j (t) is the instantaneous bit error rate of the j-th link at time t, r j (t) is the dynamic error correction code recovery efficiency of the j-th link at time t, N j (t) is the number of active data chunks of the j-th link at time t, is the instantaneous change rate of the fault propagation path length, M(t) is the real-time number of available redundant links, α, β are the system dynamic adjustment coefficients, is the instantaneous utilization rate of the q-th type of resource, C q,total is the total amount of the q-th type of resource.

[0079] The quantum topology photonic crystal optimization algorithm QTPC is used to determine the numerical values of the multi-objective optimization function indicators, and to achieve the global optimal matching of the encryption-self-healing strategy. The process is as follows:

[0080] (1) Quantum topology initialization stage: Set the range threshold and the maximum number of iterations. Based on the quantum chaotic sequence generator of the dynamic encryption module, construct a topologically protected quantum state in the periodic structure of the photonic crystal. Through the valley Hall effect, map the hierarchical characteristics of the power data to the high-symmetry path in the momentum space, and generate the initial quantum state carrying the encryption key seed. The formula is:

[0081]

[0082] Among them, ψ i (0) is the wave function of the i-th quantum topology state at the initial moment, representing a candidate solution to the optimization problem, is the quantum topology photonic crystal encoding operator, which maps the power data characteristics to the energy band structure of the photonic crystal, k m is the m-th sampling point in the momentum space, corresponding to the hierarchical characteristics of the power data at the virtual power plant terminal, r is the spatial coordinate position of the photonic crystal, χ v,K (r) is the wave function of the topological boundary state with Chern number v = 1, K is the high-symmetry point in the momentum space, providing topological protection characteristics, M is the total number of sampling points in the momentum space, which is consistent with the number of power data levels. The dynamic parameter S output by the quantum chaotic sequence generator SM9 (t) modulates the phase φ of the defect mode of the photonic crystal m , so that the key generation process forms a quantum entanglement with the topological energy band structure. The lattice constant a and the encryption intensity coefficient k i (t) satisfy Δ gap = a · k i (t), realizing the adaptive adjustment of the encryption intensity at the physical layer;

[0083] (2) Anomalous quantum Hall evolution stage, combined with the real-time link state data N j (t) and the bit error rate ∈ j (t), construct a dynamic artificial gauge field to drive the Berry curvature evolution, and achieve dissipationless optimization search through the quantum anomalous Hall effect. The specific dynamic equation is:

[0084]

[0085] Among them, e is the elementary charge quantity, is the reduced Planck constant, ε(N j (t)) is the virtual electric field strength, dynamically adjusted by the number of active data blocks N j (t) of the j-th link, Ω(k) is the Berry curvature tensor, is the momentum space gradient operator, driving the quantum state to evolve towards the low energy direction, β(∈ j (t)) is the adaptive diffusion coefficient, negatively correlated with the instantaneous bit error rate ∈ j (t) of the j-th link, is the momentum space Laplacian operator, controlling the diffusion range of the quantum state. When N j (t) decreases, the electric field is enhanced to expand the search range. When ∈ j (t) increases, the diffusion coefficient β is suppressed to prevent local convergence, forming a closed-loop balance between encryption reliability and self-healing response speed;

[0086] (3) Photo-acoustic coupling weight allocation stage. Through the strong coupling between surface plasmons and phonon modes, map the multi-objective function of the collaborative optimization module to the quantum degenerate state. The Rabi oscillation frequency is jointly determined by the power data grading parameter D i (t) and the dynamic encryption intensity coefficient k i (t), and realize the dynamic physical allocation of the weight coefficient. The formula is:

[0087]

[0088] Among them, j = 1 corresponds to the encryption delay T enc , j = 2 corresponds to the data integrity loss L int , j = 3 corresponds to the self-healing response time T heal , j = 4 corresponds to the resource utilization rate U res ; The photo-acoustic coupling strength g is positively correlated with the computing resources C q,used of the edge node, ensuring that the weight allocation process adapts to changes in resource load. w j is the weight coefficient of the j-th optimization objective, D i (t) is the instantaneous size of the i-th level data packet at time t, k i(t) The dynamic encryption strength coefficient of the i-th level data, with a value range of [0.1, 1.0], is generated by the SM9 algorithm;

[0089] (4) In the dynamic topology reconstruction stage, based on the gradient matrix W output by the edge computing node federated learning framework FL , the dielectric constant distribution of the photonic crystal is reconstructed in real time, enabling the optimization process to adapt to the topological changes of the power network. The formula is:

[0090]

[0091] where ∈0 is the base dielectric constant, ReLU is the rectified linear unit activation function, is the partial derivative of the multi-objective optimization function F with respect to the weight w j , reflecting the sensitivity of each objective. The federated gradient W synchronized by the lightweight consensus mechanism FL , after being activated by the rectified linear unit, dynamically adjusts the dielectric constant ∈(r,t), forming the co-evolution of the encryption policy library and the self-healing rule library. The topological protection energy gap width Δ gap and the resource utilization rate Ures show a negative correlation, achieving the quantization trade-off between security and efficiency;

[0092] (5) In the quantum error correction output stage, the surface code quantum error correction protocol is used to perform anti-interference coding on the optimized solution set to support the reliability of local decision-making in the offline mode;

[0093] (6) Termination condition determination: When any of the following conditions is met, the optimization iteration is terminated:

[0094] 1. Hard constraints of the power system: T enc <10ms and T heal <50ms;

[0095] 2. Quantum fidelity threshold: Tr(ρ out ρ ideal )>0.9999;

[0096] 3. Resource protection mechanism: U res >95% for 1s continuously.

[0097] Optimize the algorithm optimization process, and use the quantum tunneling effect to break through the traditional solution space limit. The formula is:

[0098]

[0099] where the tunneling coefficient k = 0.1 / ms is synchronized with the key update event to ensure the time consistency of the encryption parameters and the resource scheduling strategy, is the original solution without boundary processing, L j , U jare the lower and upper bounds of the j-th optimization variable, k is the tunneling coefficient, which controls the convergence rate when the solution vector exceeds the boundary, and T refresh is the SM9 key update timestamp, which is synchronized with the dynamic encryption module to ensure timing consistency.

[0100] Edge computing nodes deploy the federated learning framework. According to the global optimal matching strategy of the encryption-self-healing policy, they perform terminal-edge cloud collaborative training, dynamically update the encryption policy library and the self-healing rule library, support local decision-making in the offline mode, and global optimization in the online mode; the edge nodes are built-in with a lightweight consensus mechanism, and through local model gradient aggregation and global parameter synchronization, ensure the consistency of the encryption policy and the self-healing rule.

[0101] The following presents the relevant experimental data in multiple dimensions such as encryption strength, fault recovery, optimization efficiency, and anti-attack ability for the above-mentioned solution:

[0102] 1. Encryption performance comparison: This invention conducts encryption experiments on dynamic SM9 + quantum chaos, traditional SM9 static encryption, and AES-256. The test conditions are 1000 terminal nodes, a data packet size of 1MB, and a network bandwidth of 1Gbps.

[0103] Table 1: Comparison of encryption delay, key update frequency, anti-quantum attack ability, and data integrity score of different encryption methods

[0104]

[0105] It can be seen from the data in Table 1 that the dynamic SM9 + quantum chaos algorithm adopted by this invention has a high anti-two-word attack ability, greatly reduces the encryption delay, has a faster key update frequency, and has high security performance.

[0106] 2. Fault recovery performance: This invention simulates a 200-node virtual power plant network, including 5 * 0 communication links and 150 power physical links.

[0107] Table 2 shows different fault types

[0108]

[0109] 3. Algorithm optimization results: The quantum topological photonic crystal optimization algorithm QTPC adopted by this invention is used for objective optimization and compared with the genetic algorithm and the particle swarm optimization algorithm. The specific comparison results are shown in Table 3:

[0110] Table 3: Comparison of specific optimization results of different optimization algorithms

[0111]

[0112] As can be seen from the data in the table, the global optimal matching of the encryption-self-healing strategy optimized by the quantum topological photonic crystal optimization algorithm QTPC of the present invention can greatly reduce the encryption experiment, while having a small data loss and a short power grid self-healing time.

[0113] 4. Anti-attack ability

[0114] Table 4: Comparison of the defense success rate and key leakage risk between the present application and the traditional defense system under different attack types

[0115]

[0116] The traditional system refers to the existing scheme that does not adopt dynamic SM9+ quantum chaos encryption (in this embodiment, the traditional system adopts AES-256 static encryption). As can be seen from the above table, for different attack types, compared with the traditional one, the encryption transmission and self-healing control collaborative system of the present invention has an improved defense success rate, and during the operation of the system, the key leakage level is relatively low, and it has good security performance.

[0117] The above embodiments are only for explaining the technical concept and characteristics of the present invention, and their purpose is to enable those who are familiar with this technology to understand the content of the present invention and implement it accordingly, and it cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit and essence of the present invention should be covered within the protection scope of the present invention.

Claims

1. A collaborative system for encrypted transmission and self-healing control for virtual power plant terminals, characterized in that It includes a dynamic encryption module, a self-healing control module, a collaborative optimization module, and an edge computing node; The dynamic encryption module constructs a hierarchical encryption system based on quantum chaotic sequences and the national cryptographic SM9 algorithm, and generates dynamic session keys in real time through edge nodes to encrypt terminal power data in grades. The hierarchical encryption system includes: a quantum chaotic sequence generator, a key derivation function, and a multi-level encryption channel. The quantum chaotic sequence generator generates an initial entropy source based on the terminal device fingerprint and environmental noise, and generates a chaotic sequence through non-linear feedback modulation. The key derivation function performs an exclusive OR operation on the bilinear pair mapping in the SM9 algorithm and the chaotic sequence to generate a dynamic key seed, and expands the seed into a multi-level session key through the HKDF standard. The multi-level encryption channel encrypts data according to a differential encryption strategy based on the data sensitivity level; The self-healing control module integrates a blockchain distributed ledger and a deep reinforcement learning DRL model, constructs a health profile of terminal devices in real time, monitors the link status through a topology awareness algorithm, predicts the fault propagation path in combination with the Lyapunov stability criterion, and triggers communication link reconstruction and elastic scheduling of computing resources for self-healing; The collaborative optimization module constructs a multi-objective optimization function to balance the encryption strength and self-healing efficiency, and uses a quantum topology photonic crystal optimization algorithm to determine the numerical values of the multi-objective optimization function indicators to achieve the global optimal matching of the encryption-self-healing strategy. The optimization objective function includes four indicators: encryption delay, data integrity loss, self-healing response time, and resource utilization; The edge computing node deploys a federated learning framework, and performs terminal-edge cloud collaborative training according to the global optimal matching strategy of the encryption-self-healing strategy, dynamically updates the encryption policy library and the self-healing rule library, supports local decision-making in the offline mode, and global optimization in the online mode. The edge node has a lightweight consensus mechanism built in, and ensures the consistency of the encryption policy and the self-healing rule through local model gradient aggregation and global parameter synchronization.

2. The co - system of encrypted transmission and self - healing control for a virtual power plant terminal according to claim 1, characterized in that, The data sensitivity level is divided into three levels of data, and the differential encryption strategy is as follows: Level 1 data uses a composite encryption of SM9 identity encryption and chaotic sequence stream cipher; Level 2 data uses a dynamic message authentication code generated by adding a chaotic sequence after SM9 encryption; Level 3 data only uses the basic SM9 encryption, and the session key update period is synchronized with the power data collection frequency.

3. The co - system for collaborative encrypted transmission and self - healing control for a virtual power plant terminal according to claim 1, wherein, Integrate a blockchain distributed ledger and a deep reinforcement learning DRL model, and construct a health profile of terminal devices in real time, specifically as follows: 1) The blockchain distributed ledger records the voltage volatility δ of the terminal device v , the temperature drift ΔT, the communication error rate ε, the operating state parameters, and the integrity score of the encrypted link, and constructs an anti-tampering data chain through the Merkle tree; 2) The health profile is constructed based on a statistical physics model of multi-dimensional state parameters, specifically divided into: feature extraction, state dimension reduction, health calculation, and dynamic optimization; Feature extraction: Extract the mean value of time-domain features from the historical data of the device and variance The wavelet packet energy entropy E of frequency-domain features T and the Lyapunov exponent λ of non-linear features to form an 8-dimensional feature vector F(t) ∈ 8 ; State dimensionality reduction: Unsupervised dimensionality reduction of F(t) is performed through a deep autoencoder to generate a 3-dimensional latent health state Z(t) ∈ 3 ; Health degree calculation: Define the health index as: where μ z and ∑ z are the mean and covariance matrix of the potential state under normal working conditions. The DRL model driven by the dynamic optimization design double Q network dynamically optimizes the health assessment strategy with the reward function. Define the reward function as: where is the fault label function.

4. A collaborative system for encrypted transmission and self-healing control for a virtual power plant terminal according to claim 1, characterized in that, Monitor the link status through a topology awareness algorithm, predict the fault propagation path in combination with the Lyapunov stability criterion, and trigger communication link reconstruction and elastic scheduling of computing resources for self-healing, specifically as follows: 1) The topology-aware algorithm monitors power communication links and physical power links in real time and constructs a cross-layer link state matrix. The power communication link is the blockchain data transmission channel, and the parameters include the bit error rate ∈ j (t), the dynamic error correction code recovery efficiency r j (t), the number of active data chunks N j (t), the time delay τ j (t), and the bandwidth B j (t); the physical power link is the power equipment connection line, and the parameters include the line power flow P j (t), the voltage volatility δ v (t), the temperature drift ΔT(t), and the on / off state C j (t) between nodes. The cross-layer link state matrix Q(t) ∈ m×5 is: where m is the total number of links, and j = 1, 2, … m; 2) The Lyapunov stability predicts the fault propagation path of the power physical line and the fault propagation path of the communication link; According to the power flow equation of the power line Calculate the fault propagation path length L path (t). When detected , trigger the communication link reconstruction based on SDN, and switch to the redundant path within T reconfig < 50 ms and calculate the elastic resource scheduling. Allocate the edge node resources dynamically according to C used (t) = min(1.2C base , C total ). If the physical link C j (t) = 0, then synchronously increase the encryption strength coefficient k i (t) ≥ 0.8 and start data retransmission, where T reconfig is the communication link reconstruction time, C base is the reference computing resource requirement, and C total is the total available computing resource.

5. The encryption transmission and self-healing control collaborative system for a virtual power plant terminal according to claim 4, characterized in that, The multi-objective optimization function F of the collaborative optimization module is specifically as follows: f = [f1, f2, f3, f4] T where \(f = [f_1, f_2, f_3, f_4]\) T is the factor vector, \(w = [w_1, w_2, w_3, w_4]\) T is the weight vector, ° represents the Hadamard product, and the weights are dynamically allocated through the quantum topological photonic crystal optimization algorithm. \(f_1\) is the encryption delay efficiency factor, \(f_2\) is the data integrity factor, \(f_3\) is the self-healing response efficiency factor, \(f_4\) is the resource utilization efficiency factor, \(w_1\) is the weight coefficient of the encryption delay, \(w_2\) is the weight coefficient of the data integrity loss, \(w_3\) is the weight coefficient of the self-healing response time, \(w_4\) is the weight coefficient of the resource utilization rate, \(T\) enc is the encryption delay, \(L\) int is the data integrity loss, \(T\) heal is the self-healing response time, \(U\) res is the resource utilization rate, is the encryption delay reference value, is the data integrity reference value, \(U\) base is the resource utilization rate reference value, \(D\) i is the size of the \(i\)-th level data packet, \(v\) max is the maximum encryption rate, \(\in\) max is the maximum allowable bit error rate, \(N\) max is the maximum number of active data chunks, \(C\) total is the total computing resource, \(p\) is the total number of resource types, \(D\) i \((t)\) is the instantaneous size of the \(i\)-th level data packet at time \(t\), \(k\) i \((t)\) is the dynamic encryption intensity coefficient, \(v(t)\) is the real-time encryption rate of the edge node, \(\Delta t\) is the time window of the encryption session, \(\in\) j \((t)\) is the instantaneous bit error rate of the \(j\)-th link at time \(t\), \(r\) j \((t)\) is the dynamic error correction code recovery efficiency of the \(j\)-th link at time \(t\), \(N\) j \((t)\) is the number of active data chunks of the \(j\)-th link at time \(t\), is the instantaneous change rate of the fault propagation path length, \(M(t)\) is the real-time number of available redundant links, \(\alpha, \beta\) are system dynamic adjustment coefficients, is the instantaneous utilization rate of the \(q\)-th type of resource, \(C\) q,total is the total amount of the \(q\)-th type of resource.

6. The collaborative system for encrypted transmission and self-healing control for a virtual power plant terminal according to claim 5, characterized in that, The quantum topological photonic crystal optimization algorithm QTPC is used to determine the numerical values of the multi-objective optimization function indicators, and to achieve the global optimal matching of the encryption-self-healing strategy. The process is as follows: (1) In the quantum topology initialization stage, the range threshold and the maximum number of iterations are set. Based on the quantum chaotic sequence generator of the dynamic encryption module, a topologically protected quantum state is constructed in the periodic structure of the photonic crystal. Through the valley Hall effect, the hierarchical characteristics of the power data are mapped to the high-symmetry path in the momentum space, and the initial quantum state carrying the encryption key seed is generated. The formula is as follows: Among them, ψ i (0) is the wave function of the i-th quantum topological state at the initial moment, representing a candidate solution to the optimization problem. is the quantum topological photonic crystal encoding operator, which maps the power data characteristics to the photonic crystal energy band structure. k m is the m-th sampling point in the momentum space, corresponding to the hierarchical characteristics of the power data at the virtual power plant terminal. r is the spatial coordinate position of the photonic crystal. χ v,K (r) is the wave function of the topological boundary state with Chern number v = 1. K is the high-symmetry point in the momentum space, providing topological protection characteristics. M is the total number of sampling points in the momentum space, which is consistent with the number of power data levels. The dynamic parameter S SM9 (t) modulates the phase φ m of the defect mode of the photonic crystal, so that the key generation process forms quantum entanglement with the topological energy band structure. The lattice constant a and the encryption intensity coefficient k i (t) satisfy Δ gap = a·k i (t), realizing the adaptive adjustment of the encryption intensity at the physical layer. (2) Anomalous quantum Hall evolution stage, combined with the real-time link state data N of the self-healing control module j (t) and the bit error rate ∈ j (t), construct a dynamic artificial gauge field to drive the Berry curvature evolution, and achieve dissipationless optimization search through the quantum anomalous Hall effect. The specific dynamic equation is as follows: where \(e\) is the elementary charge, is the reduced Planck constant, \(\epsilon(N j (t))\) is the virtual electric field strength, which is dynamically adjusted by the number of active data chunks \(N j (t)\) of the \(j\)-th link. \(\Omega(k)\) is the Berry curvature tensor, is the momentum space gradient operator, driving the quantum state to evolve in the direction of lower energy. \(\beta(\in j (t))\) is the adaptive diffusion coefficient, which is negatively correlated with the instantaneous bit error rate \(\in j (t)\) of the \(j\)-th link, is the momentum space Laplacian operator, controlling the diffusion range of the quantum state. When \(N j (t)\) decreases, the electric field is enhanced to expand the search range. When \(\in j (t)\) increases, the diffusion coefficient \(\beta\) is suppressed to prevent local convergence, forming a closed-loop balance between encryption reliability and self-healing response speed; (3) In the light-acoustic coupling weight allocation stage, through the strong coupling between surface plasmons and phonon modes, the multi-objective function of the collaborative optimization module is mapped to the quantum degenerate state, and the Rabi oscillation frequency is determined by the power data classification parameter D i (t) and the dynamic encryption intensity coefficient k i (t) jointly determine to achieve the dynamic physical allocation of the weight coefficient. The formula is: Among them, j = 1 corresponds to the encryption delay T enc , j = 2 corresponds to the data integrity loss L int , j = 3 corresponds to the self-healing response time T heal , j = 4 corresponds to the resource utilization rate U res ; The opto-acoustic coupling intensity g is positively correlated with the computing resources C of the edge node q,used , ensuring that the weight allocation process adapts to changes in resource load, w j is the weight coefficient of the j-th optimization objective, D i (t) is the instantaneous size of the i-th level data packet at time t, k i (t) is the dynamic encryption intensity coefficient of the i-th level data, and its value range is [0.1, 1.0], which is generated by the SM9 algorithm; (4) During the dynamic topology reconstruction stage, based on the gradient matrix W output by the edge computing node federated learning framework FL , the dielectric constant distribution of the photonic crystal is reconstructed in real time, enabling the optimization process to adapt to the topology changes of the power network. The formula is as follows: where ∈0 is the fundamental permittivity, ReLU is the rectified linear unit activation function, is the partial derivative of the multi-objective optimization function F with respect to the weight w j , reflecting the sensitivity of each objective, and the federated gradient W synchronized by the lightweight consensus mechanism FL After being activated by the rectified linear unit, dynamically adjusts the permittivity ∈(r,t), forming the co-evolution of the encryption policy library and the self-healing rule library, and the topological protection energy gap width Δ gap has a negative correlation with the resource utilization rate Ures, realizing the quantization trade-off between security and efficiency; (5) In the quantum error correction output stage, the surface code quantum error correction protocol is used to perform anti-interference coding on the optimized solution set to support the reliability of local decision-making in the offline mode; (6) Judgment of termination conditions: The optimization iteration is terminated when any of the following conditions is met: 1). Hard constraints of the power system: T enc < 10 ms and T heal < 50 ms; 2). Quantum fidelity threshold: Tr(ρ out ρ ideal ) > 0.9999; 3). Resource protection mechanism: U res > 95% has a continuous timeout of 1 s.

7. A collaborative system for encrypted transmission and self-healing control for a virtual power plant terminal according to claim 6, characterized in that, Optimize the algorithm optimization process, and use the quantum tunneling effect to break through the traditional solution space limit. The formula is as follows: Among them, the tunneling coefficient \(k = 0.1 / ms\) is synchronized with the key update event to ensure the time consistency of the encryption parameters and the resource scheduling strategy. is the original solution without boundary processing, \(L\) j , \(U\) j are the lower and upper bounds of the \(j\)-th optimization variable, \(k\) is the tunneling coefficient, which controls the convergence speed when the solution vector exceeds the boundary, \(T\) refresh is the SM9 key update timestamp, which is synchronized with the dynamic encryption module to ensure timing consistency.

Images