Dynamic reconstruction and self-healing control method and system for expressway energy network

Abstract

The invention provides an expressway energy network dynamic reconstruction and self-healing control method and system, and the method comprises the steps: obtaining the state information of each node in an expressway energy network in real time, and employing a deep learning algorithm, optimizing a power distribution strategy of the expressway energy network by learning the optimal action of each node in different states; defining the expressway energy network as a graph structure, and optimizing a power transmission path of the expressway energy network by adopting a graph neural network; dynamically adjusting the network topology of the expressway energy network by adopting a particle swarm optimization algorithm so as to optimize power distribution and a transmission path; constructing and training an expressway energy network fault prediction model by using a deep neural network, and predicting the fault of the expressway energy network by using the expressway energy network fault prediction model; and when it is predicted that the expressway energy network has a fault, optimizing a self-healing strategy of the expressway energy network by using DRL, and executing the self-healing strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of power system dispatching automation, and more specifically, to a method and system for dynamic reconfiguration and self-healing control of a highway energy network. Background Art

[0002] With the transformation of the global energy structure and the rapid development of green energy, the highway energy system has gradually become an important part of modern transportation infrastructure. In recent years, with the popularization of electric vehicles and the wide application of renewable energy (such as photovoltaic and wind power) along highways, the demand for highway energy networks has become increasingly urgent. However, traditional AC power supply systems face great challenges in terms of distributed energy access along highways, load demand fluctuations, and energy dispatching flexibility, and it is difficult to meet the requirements of future green transportation.

[0003] Specifically, when applying traditional AC power supply systems to current highway energy networks, the following key problems exist: 1. Uneven distribution: Loads and distributed energy access points along highways are usually uneven, resulting in energy surplus or shortage in local areas, and traditional power transmission methods are difficult to dynamically adjust power distribution.

[0004] 2. Energy fluctuations: Distributed renewable energy (such as photovoltaic and wind power) accessed along highways is greatly affected by natural conditions and has strong volatility. This volatility poses a great challenge to the stability of the power grid, especially when dealing with rapid changes in load demand.

[0005] 3. Energy loss: During long-distance power transmission, the loss of traditional AC systems is large, especially the transmission efficiency of distributed energy is low, resulting in energy waste.

[0006] 4. Insufficient system reliability and self-healing ability: Traditional highway energy networks lack flexible control means and self-healing ability. When a fault occurs, it is easy to cause partial or entire system shutdown, affecting traffic safety and the stability of power supply. Summary of the Invention

[0007] Aiming at the above problems, the purpose of the present invention is to provide a method and system for dynamic reconfiguration and self-healing control of a highway energy network, which uses flexible DC technology combined with cutting-edge algorithms such as deep reinforcement learning, graph neural network, and particle swarm optimization to achieve intelligent dispatching, path optimization, ring network dynamic reconfiguration, and self-healing control of the highway energy network, and improve the reliability, flexibility, and energy utilization efficiency of the highway energy network.

[0008] To achieve the above object, the present invention is realized through the following technical solutions: In the first aspect, the present invention discloses a method for dynamic reconstruction and self-healing control of a highway energy network, including: Obtain the status information of each node in the highway energy network in real time, and adopt a deep learning algorithm to optimize the power distribution strategy of the highway energy network by learning the optimal actions of each node in different states; Based on the structural information of the highway energy network, define the highway energy network as a graph structure, and use a graph neural network to optimize the power transmission path of the highway energy network; Adopt a particle swarm optimization algorithm to dynamically adjust the network topology of the highway energy network to optimize power distribution and transmission paths; Use a deep neural network to construct and train a fault prediction model for the highway energy network, and based on the real-time status information of each node in the highway energy network, use the fault prediction model of the highway energy network to predict the faults of the highway energy network; When it is predicted that there are faults in the highway energy network, use DRL to optimize the self-healing strategy of the highway energy network and execute the self-healing strategy.

[0009] Further, the obtaining the status information of each node in the highway energy network in real time, and adopting a deep learning algorithm to optimize the power distribution strategy of the highway energy network by learning the optimal actions of each node in different states includes: Obtain the voltage V of each node in the highway energy network in real time i , current I i , power demand P load,i ; Define a state space S and an action space A for the highway energy network; Set a reward function R and an update formula for the action value function Q(s,a) for the deep learning algorithm; Based on the set deep learning algorithm, optimize the power distribution strategy of the highway energy network. Further, the highway energy network includes n nodes, and the state space S of the highway energy network is: S={s 1 , s 2 , …, s n}; Among them, the state of node i is defined as: s i = [V i , I i , P load,i ; The action space A is defined as the power distribution decision made for each node; for m nodes, the action set is: A={a 1 , a 2,…,a m}; where each action a i represents the power allocation value P for the i-th node, i.e.: a i = P i ; alloc,I ; The reward function R is:

[0010] where P loss,i is the power loss of node i, P alloc,i is the power allocated to node i, P load,i is the power demand of node i, and α is a penalty coefficient used to balance the effects of loss and load imbalance; The update formula for the action-value function Q(s,a) is specifically as follows:

[0011] where: η is the learning rate, is the reward at the current time step, γ is the discount factor, is the Q value of the current state-action pair; is the next state 's maximum Q value. Furthermore, based on the structural information of the highway energy network, the highway energy network is defined as a graph structure, and a graph neural network is used to optimize the power transmission path of the highway energy network, including: Represent the highway energy network as a graph G=(V,E); where V is the set of nodes in the highway energy network, representing each power node in the power grid; E is the power transmission line between power nodes; Define that each power node v i has a feature vector x i ; x i = [V i , I i , P load,i ; where V i , I i , P load,i are the voltage, current, and load of the power node v i , respectively; Set the update formula for the hidden state of the power node v i in the graph neural network at the t-th round as:

[0012] where, is the power node v i at the hidden state in the t-th round; W h and W m are weight matrices, and σ is a non-linear activation function; Set the total loss function L of the graph neural network as:

[0013] where P loss,ij is the power loss between the power node v i and v j ; T ij is the transmission time of electric energy between the power node v i and v j , and β is the time penalty coefficient.

[0014] Furthermore, the particle swarm optimization algorithm is used to dynamically adjust the network topology of the highway energy network to optimize power distribution and transmission paths, including: In the particle swarm optimization algorithm, N particles are set, each particle represents a network configuration, and the initial position of each particle is x i (0), and the initial velocity is v i (0); In the t-th round of iteration, the position and velocity update formulas of the particle are as follows:

[0015] where ω is the inertia weight, c 1 , c 2 are acceleration coefficients, r 1 , r 2 are random numbers, p i is the historical optimal position of particle i, and g is the global optimal position; Define the fitness function f(x i ) for the particle swarm optimization algorithm, specifically as follows:

[0016] where P loss,i is the power loss of node i, λ is the fault impact penalty coefficient, and FaultImpact(i) represents the degree to which node i is affected by the fault; Based on the configured particle swarm optimization algorithm, dynamically adjust the network topology of the highway energy network. Furthermore, set a loss function for the highway energy network fault prediction model to measure the accuracy of fault prediction; The loss function is specifically as follows:

[0017] Among them, is the actual label, is the predicted value.

[0018] Furthermore, the self-healing strategy for optimizing the highway energy network using DRL includes: Define the fault state as S fault and the restoration action as A recover . The adjustment control parameter of the highway energy network at each moment is θ; Set the update formula of the control parameter θ as:

[0019] where η is the learning rate.

[0020] In a second aspect, the present invention also discloses a dynamic reconfiguration and self-healing control system for a highway energy network, including: A scheduling optimization module, which is used to obtain the state information of each node in the highway energy network in real time, adopt a deep learning algorithm, and optimize the power distribution strategy of the highway energy network by learning the optimal actions of each node in different states; A power transmission path optimization module, which is used to define the highway energy network as a graph structure based on the structural information of the highway energy network, and optimize the power transmission path of the highway energy network using a graph neural network; A ring network dynamic reconfiguration module, which is used to dynamically adjust the network topology of the highway energy network using a particle swarm optimization algorithm to optimize power distribution and transmission paths; A fault prediction module, which is used to build and train a highway energy network fault prediction model using a deep neural network, and predict the faults of the highway energy network based on the real-time state information of each node in the highway energy network; A ring network self-healing control module, which is used to optimize the self-healing strategy of the highway energy network using DRL and execute the self-healing strategy when it is predicted that there are faults in the highway energy network.

[0021] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. By applying the flexible DC transmission technology, the present invention optimizes the power transmission path of the highway energy network, reduces energy losses, especially for long-distance transmission and multi-terminal interconnection, improves the power transmission efficiency, and realizes efficient power scheduling between different regions.

[0022] 2. Aiming at the uneven problems of load distribution and renewable energy access along the highway, the present invention realizes multi-terminal interconnection and flexible control by reasonably arranging DC power nodes, can dynamically adjust the grid topology according to real-time needs, and ensures the balanced distribution and stable supply of power.

[0023] 3. The present invention utilizes distributed control technology and advanced optimization algorithms to dynamically adjust the power transmission path and node control parameters in response to the volatility of distributed renewable energy sources such as photovoltaic and wind power along highways, reducing the impact of fluctuations on grid stability and enhancing the anti-disturbance ability of the energy network.

[0024] 4. In the event of a failure in the highway energy network or the failure of an energy node, the present invention can automatically identify the fault and quickly restore system operation by dynamically reconstructing the ring network structure. The self-healing strategy is optimized using deep reinforcement learning algorithms, enabling the system to have stronger fault tolerance and self-repair capabilities, and enhancing the overall stability and security of the system.

[0025] 5. The present invention can dynamically adjust the control parameters of the system based on the real-time monitored ring network status and topological changes, optimize the allocation of power resources, and ensure that the system always operates in an optimal state under different load and fault conditions, maximizing the stability and efficiency of the system.

[0026] Therefore, compared with the prior art, the present invention has outstanding substantive features and significant progress, and the beneficial effects of its implementation are also obvious. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings according to the provided drawings without creative efforts.

[0028] Figure 1 It is a flowchart of the method of the specific implementation manner of the present invention.

[0029] Figure 2 It is a system structure diagram of the specific implementation manner of the present invention.

[0030] In the figure, 1. Dispatching optimization module; 2. Power transmission path optimization module; 3. Ring network dynamic reconstruction module; 4. Fault prediction module; 5. Ring network self-healing control module. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0031] In order to enable those skilled in the art to better understand the solution of the present invention, the present invention will be further described in detail below with reference to the drawings and specific implementation manners. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0032] See Figure 1 As shown, this embodiment provides a method for dynamic reconfiguration and self-healing control of a highway energy network, including the following steps: S1: Real-time obtain the state information of each node in the highway energy network, and adopt a deep learning algorithm to optimize the power distribution strategy of the highway energy network by learning the optimal actions of each node in different states.

[0033] In the specific implementation, first, real-time obtain the voltage V i , current I i , and power demand P load,i of each node in the highway energy network. Then define the state space S and action space A for the highway energy network; At the same time, set the reward function R and the update formula of the action value function Q(s,a) for the deep learning algorithm. Finally, based on the set deep learning algorithm, optimize the power distribution strategy of the highway energy network.

[0034] Among them, the highway energy network includes n nodes, and the state space S of the highway energy network is: S={s 1 ,s 2 ,…,s n}; The state of node i is defined as: s i =[V i ,I i ,P load,i .

[0035] The action space A is defined as the power distribution decision for each node; for m nodes, the action set is: A={a 1 ,a 2 ,…,a m}; where each action a i represents the power distribution value P i for the i-th node, that is: a i =P alloc,I .

[0036] In this step, by defining the reward function R, the efficiency of the scheduling strategy is measured, and the goal is to minimize the system power loss and load imbalance. The reward function is:

[0037] Among them, P loss,i is the power loss of node i, P alloc,i is the power allocated to node i, P load,i is the power demand of node i, and α is a penalty coefficient used to balance the effects of loss and load imbalance.

[0038] In this step, a Deep Q-Network (DQN) is used to update the action-value function Q(s,a). The update formula is:

[0039] where: η is the learning rate; is the reward at the current time step; γ is the discount factor, used to control the impact of future rewards; is the Q-value of the current state-action pair; is the next state is the maximum Q-value of. By continuously iterating, an optimal power allocation strategy can be found under different states to maximize the power utilization efficiency.

[0040] S2: Based on the structural information of the highway energy network, the highway energy network is defined as a graph structure, and a graph neural network is used to optimize the power transmission path of the highway energy network.

[0041] In the specific implementation, this step uses a Graph Neural Network (GNN) to optimize the power transmission path. Among them, the graph structure is used to represent the node and edge relationships of the power network, and the optimal path selection is realized through the message passing mechanism.

[0042] First, represent the highway energy network as a graph G=(V,E); where, V is the set of nodes in the highway energy network, representing each power node in the power grid; E is the power transmission line between power nodes.

[0043] In the graph, each node v i has a feature vector x i , which includes the voltage V i , current I i and load P load,i : x i =[V i ,I i ,P load,i .

[0044] In the graph neural network, each power node v i receives messages from its neighbor nodes N(i) and updates its hidden state through an aggregation function. The hidden state update formula of power node v i in the t-th round is:

[0045] where, is the hidden state of power node vi in the t-th round; and is the weight matrix, and σ is the non - linear activation function.

[0046] In the graph neural network, through multiple rounds of message passing, the power nodes learn the network structure around them and finally achieve path optimization. To achieve path optimization, it is necessary to define the total loss function L, and the goal is to minimize the power loss and transmission time of the network.

[0047] Set the total loss function L of the graph neural network as:

[0048] where P loss,ij is the power loss between power nodes v i and v j ; T ij is the transmission time of electric energy between power nodes v i and v j . β is the time penalty coefficient, which is used to balance the power loss and transmission time.

[0049] S3: Adopt the particle swarm optimization algorithm to dynamically adjust the network topology of the highway energy network to optimize power distribution and transmission paths.

[0050] In the specific implementation, to achieve the dynamic reconfiguration of the ring network, we adopted the particle swarm optimization algorithm (Particle Swarm Optimization, PSO). This algorithm is used to dynamically adjust the network topology, optimize power distribution and transmission paths. Specifically as follows: 1. Particle initialization: In PSO, each particle represents a possible network configuration. Assume that the system has N particles, and the initial position of each particle is x i (0), and the initial velocity is v i (0). In the t - th round of iteration, the update formulas for the position and velocity of the particle are as follows:

[0051] where ω is the inertia weight, which is used to control the influence of the current velocity; c 1 , c 2 are the acceleration coefficients, which are used to control the influence of the individual and global optimal solutions; r 1 , r 2 are random numbers, which are used to increase the diversity of the search space; p i is the historical optimal position of particle i, and g is the global optimal position.

[0052] 2. Define the fitness function f(x i ) for the particle swarm optimization algorithm, which is used to evaluate the configuration effect of each particle, and the goal is to minimize the power loss and the impact of faults. Specifically as follows:

[0053] Among them, P loss,i is the power loss of node i, λ is the fault impact penalty coefficient, and FaultImpact(i) represents the degree to which node i is affected by the fault.

[0054] Based on the configured particle swarm optimization algorithm, through multiple iterations, the particle swarm will converge to the optimal network topology configuration, thereby achieving optimal energy transmission and network self-healing.

[0055] S4: Use a deep neural network to construct and train a fault prediction model for the highway energy network. Based on the real-time status information of each node in the highway energy network, use the fault prediction model of the highway energy network to predict the faults of the highway energy network.

[0056] In the specific implementation, use a deep neural network (DNN) to construct a fault prediction model for the highway energy network, and use the historical status information of each node in the highway energy network to train the fault prediction model of the highway energy network. Then, use the fault prediction model of the highway energy network to predict the faults of the highway energy network.

[0057] Among them, when using the fault prediction model of the highway energy network for fault prediction, the loss function used is binary cross-entropy loss, which is used to measure the accuracy of fault prediction. Specifically as follows:

[0058] Among them, is the actual label (whether a fault occurs), is the predicted value.

[0059] S5: When it is predicted that there are faults in the highway energy network, use DRL to optimize the self-healing strategy of the highway energy network and execute the self-healing strategy.

[0060] In the specific implementation, when a fault is detected, use DRL to optimize the self-healing strategy.

[0061] First, define the fault state as S fault , the recovery action as A recover , and the adjustment control parameter of the highway energy network at each moment as θ; Based on this, set the update formula of the control parameter θ as:

[0062] Among them, η is the learning rate, and the goal is to minimize the fault impact and restore the normal operation of the system.

[0063] Through the above definitions and settings, this step can achieve self-healing after a fault occurs by combining deep reinforcement learning (DRL).

[0064] The present invention provides a method for dynamic reconfiguration and self-healing control of a highway energy network. By reasonably arranging flexible DC power nodes and optimizing the power transmission path, it supports multi-terminal power exchange within the highway energy network; this method can dynamically adjust the loop network topology according to the real-time grid state and load demand, optimize the allocation of power resources, and ensure the efficient operation of the grid; when a fault occurs, this method can automatically identify the fault point and quickly reconstruct the power loop network according to the fault type and scope, restore normal power supply, and improve the anti-interference ability and fault tolerance of the highway energy network.

[0065] See Figure 2 As shown, the present invention also discloses a dynamic reconfiguration and self-healing control system for a highway energy network, including: a scheduling optimization module 1, a power transmission path optimization module 2, a loop network dynamic reconfiguration module 3, a fault prediction module 4, and a loop network self-healing control module 5. The scheduling optimization module 1 is used to obtain the state information of each node in the highway energy network in real time, and adopt a deep learning algorithm to optimize the power distribution strategy of the highway energy network by learning the optimal actions of each node in different states.

[0066] The power transmission path optimization module 2 is used to define the highway energy network as a graph structure based on the structural information of the highway energy network, and adopt a graph neural network to optimize the power transmission path of the highway energy network.

[0067] The loop network dynamic reconfiguration module 3 is used to dynamically adjust the network topology of the highway energy network by using a particle swarm optimization algorithm to optimize the power distribution and transmission path.

[0068] The fault prediction module 4 is used to construct and train a highway energy network fault prediction model using a deep neural network, and predict the faults of the highway energy network based on the real-time state information of each node in the highway energy network by using the highway energy network fault prediction model.

[0069] The loop network self-healing control module 5 is used to optimize the self-healing strategy of the highway energy network using DRL and execute the self-healing strategy when it is predicted that there is a fault in the highway energy network.

[0070] The specific implementation manner of the dynamic reconfiguration and self-healing control system of the highway energy network in this embodiment is basically the same as that of the above-mentioned dynamic reconfiguration and self-healing control method of the highway energy network, and will not be elaborated here.

[0071] In summary, the present invention utilizes flexible DC technology combined with cutting-edge algorithms such as deep reinforcement learning, graph neural network, and particle swarm optimization to achieve intelligent scheduling, path optimization, ring network dynamic reconstruction, and self-healing control of the highway energy network, improving the reliability, flexibility, and energy utilization efficiency of the highway energy network.

[0072] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. For the same or similar parts among the various embodiments, reference can be made to each other. For the methods disclosed in the embodiments, since they correspond to the systems disclosed in the embodiments, the description is relatively simple. For related parts, reference can be made to the description in the method part.

[0073] Those skilled in the art can further realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0074] In several embodiments provided by the present invention, it should be understood that the disclosed systems, systems, and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of the systems or units can be in electrical, mechanical, or other forms.

[0075] The units described as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place, or they can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0076] In addition, the functional modules in the various embodiments of the present invention can be integrated in a processing unit, or each module can exist physically alone, or two or more modules can be integrated in one unit.

[0077] Similarly, in each embodiment of the present invention, each processing unit may be integrated into a functional module, may exist physically, or two or more processing units may be integrated into a functional module.

[0078] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be implemented directly in hardware, in a software module executed by a processor, or in a combination thereof. The software module may be disposed in a random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium well known in the art.

[0079] Finally, it should also be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, article or device comprising the element.

[0080] The above has introduced in detail the method and system for dynamic reconstruction and self-healing control of the highway energy network provided by the present invention. Specific examples are used herein to illustrate the principle and implementation manner of the present invention. The description of the above embodiments is only used to help understand the method and its core idea of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the protection scope of the present invention.

Claims

1. A method for dynamic reconstruction and self-healing control of a highway energy network, characterized in that: include: Obtain the status information of each node in the highway energy network in real time, and use deep learning algorithms to optimize the power allocation strategy of the highway energy network by learning the optimal actions of each node in different states; Based on the structural information of the highway energy network, the highway energy network is defined as a graph structure, and a graph neural network is used to optimize the power transmission path of the highway energy network. The particle swarm optimization algorithm is used to dynamically adjust the network topology of the highway energy grid to optimize power distribution and transmission paths; Use deep neural networks to build and train a highway energy network fault prediction model. Based on the real-time status information of each node in the highway energy network, the highway energy network fault prediction model is used to predict highway energy network faults. When a fault is predicted in the highway energy network, the self-healing strategy of the highway energy network is optimized using the DRL and executed.

2. The method for dynamic reconstruction and self-healing control of highway energy network according to claim 1 is characterized in that: The real-time acquisition of the status information of each node in the highway energy network and the use of a deep learning algorithm to optimize the power allocation strategy of the highway energy network by learning the optimal action of each node in different states include: Obtain the voltage V of each node in the highway energy network in real time i 、Current I i , power demand P load,i ; Define the state space S and action space A for the highway energy network; Set the update formula of reward function R and action value function Q(s,a) for deep learning algorithm; Based on the settings, a deep learning algorithm is adopted to optimize the power allocation strategy of the highway energy network.

3. The highway energy network dynamic reconstruction and self-healing control method according to claim 1 is characterized in that: The highway energy network includes n nodes, and the state space S of the highway energy network is: S={s1,s2,…,s n }; Among them, the state of node i is defined as: s i =[V i ,I i ,P load,i ]; The action space A is defined as the power allocation decision made for each node; for m nodes, the action set is: A={a1,a2,…,a m }; Each action a i Represents the power allocation value P for the i-th node i , that is: a i =P alloc,I ; The reward function R is: Among them, P loss,i is the power loss of node i, P alloc,i is the power allocated to node i, P load,i is the power demand of node i, α is the penalty coefficient used to balance the impact of loss and load imbalance; The update formula of the action value function Q(s,a) is as follows: Where η is the learning rate, is the reward at the current time step, γ is the discount factor, is the Q-value of the current state-action pair; is the next state The maximum Q value of .

4. The method for dynamic reconstruction and self-healing control of highway energy network according to claim 1 is characterized in that: Based on the structural information of the highway energy network, the highway energy network is defined as a graph structure, and a graph neural network is used to optimize the power transmission path of the highway energy network, including: The highway energy network is represented as a graph G=(V,E); Among them, V is the node set in the highway energy network, representing each power node in the power grid; E is the power transmission line between power nodes; Define each power node v i With the eigenvector x i ; x i =[V i ,I i ,P load,i ]; Among them, V i ,I i , P load,i , respectively, are power nodes v i voltage, current and load; The power node v in the graph neural network i The hidden state update formula in round t is set to: in, is the power node v i The hidden state at round t; W h and W m is the weight matrix, σ is the nonlinear activation function; The total loss function L of the graph neural network is set to: Among them, P loss,ij is the power node v i and v j Power loss between ij is the electrical energy at the power node v i and v j The transmission time between them is β, and β is the time penalty coefficient.

5. The method for dynamic reconstruction and self-healing control of highway energy network according to claim 1 is characterized in that: The method of dynamically adjusting the network topology of the highway energy network by using a particle swarm optimization algorithm to optimize power distribution and transmission paths includes: In the particle swarm optimization algorithm, there are N particles, each particle represents a network configuration, and the initial position of each particle is x i (0), the initial velocity is v i (0); In the tth iteration, the particle position and velocity update formula are as follows: Among them, ω is the inertia weight, c1 and c2 are acceleration coefficients, r1 and r2 are random numbers, and p i is the historical optimal position of particle i, and g is the global optimal position; Define the fitness function f(x i ), as follows: Among them, P loss,i is the power loss of node i, λ is the fault impact penalty coefficient, and FaultImpact(i) indicates the degree to which node i is affected by the fault; Based on the configured particle swarm optimization algorithm, the network topology of the highway energy grid is dynamically adjusted.

6. The highway energy network dynamic reconstruction and self-healing control method according to claim 5 is characterized in that: Setting a loss function for the highway energy network fault prediction model to measure the accuracy of fault prediction; The loss function is as follows: in, is the actual label, is the predicted value.

7. The method for dynamic reconstruction and self-healing control of highway energy network according to claim 6 is characterized in that: The self-healing strategy of using DRL to optimize the highway energy network includes: Define the fault state as S fault , restore action to A recover , the adjustment control parameter of the highway energy network at each moment is θ; The update formula of the control parameter θ is set as: Where η is the learning rate.

8. A highway energy network dynamic reconstruction and self-healing control system, characterized in that: The system adopts the highway energy network dynamic reconstruction and self-healing control method as described in any one of claims 1 to 7; The system comprises: The scheduling optimization module is used to obtain the status information of each node in the highway energy network in real time, and uses a deep learning algorithm to optimize the power allocation strategy of the highway energy network by learning the optimal action of each node under different states; The power transmission path optimization module is used to define the highway energy network as a graph structure based on the structural information of the highway energy network, and use a graph neural network to optimize the power transmission path of the highway energy network; The ring network dynamic reconstruction module is used to dynamically adjust the network topology of the highway energy network using a particle swarm optimization algorithm to optimize power distribution and transmission paths; A fault prediction module is used to build and train a highway energy network fault prediction model using a deep neural network, and to predict highway energy network faults using the highway energy network fault prediction model based on the real-time status information of each node in the highway energy network; The ring network self-healing control module is used to optimize the self-healing strategy of the highway energy network using the DRL and execute the self-healing strategy when a fault is predicted in the highway energy network.

Images