Charging pile intelligent control system with multiple safety protection

Abstract

The invention relates to the technical field of charging pile intelligent control, and particularly discloses a charging pile intelligent control system with multiple safety protections, which is provided with a charging pile multi-modal data acquisition module, a monitoring risk matrix construction module, a game strengthening control strategy generation module and a charging pile defense network diagram output module. The problems that traditional safety monitoring depends on various sensors, only specific parameter detection can be carried out, complex risk scenes faced by the charging piles are difficult to fully cover, the capacity of monitoring some comprehensive faults or potential hidden dangers is limited, and along with the increase of the number of the charging piles, a traditional management mode is difficult to meet operation and maintenance requirements are solved. The method is beneficial for optimizing resource allocation, enhancing the capability of the charging pile system for resisting external attacks and internal faults, and can dynamically adjust and adapt to complex and variable operation scenes according to real-time change conditions, so that stable and efficient operation of the charging pile is guaranteed, and the friendliness to a power grid is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent control of charging piles, and in particular to an intelligent control system of charging piles with multiple safety protections. Background Art

[0002] With the increasing global awareness of environmental protection and the need for energy structure transformation, electric vehicles, as a clean and efficient means of transportation, have developed rapidly in recent years. The widespread use of electric vehicles is inseparable from the support of charging piles, a core infrastructure. The market share of charging piles has risen rapidly, which has led to the continuous expansion of the construction and layout of charging piles. During the long-term operation of charging piles, electrical components age and mechanical structures wear out from time to time, which may cause faults such as overcurrent, overvoltage, and short circuit, affecting the normal operation of the equipment and even causing serious accidents such as fires.

[0003] Nowadays, there are some deficiencies in the research on intelligent control of charging piles. Specifically, the traditional reliance on various sensors for safety monitoring can only detect specific parameters, making it difficult to fully cover the complex risk scenarios faced by charging piles, and the monitoring capabilities for some comprehensive faults or potential hidden dangers are limited. With the increase in the number of charging piles, traditional management methods are difficult to meet the needs of operation and maintenance. Intelligent control systems can realize functions such as real-time monitoring, fault warning, and remote control to improve operation and maintenance efficiency. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a charging pile intelligent control system with multiple safety protections, which can effectively solve the problems involved in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a charging pile intelligent control system with multiple safety protections, including a charging pile multimodal data acquisition module, a monitoring risk matrix construction module, a game-enhanced control strategy generation module and a charging pile defense network diagram output module: the charging pile multimodal data acquisition module collects charging pile intelligent monitoring multimodal data and outputs the original charging pile intelligent monitoring status set; the monitoring risk matrix construction module constructs a charging pile intelligent monitoring risk matrix based on the original charging pile intelligent monitoring status set; the game-enhanced control strategy generation module is based on the charging pile intelligent monitoring risk matrix, and obtains power grid power, deep deterministic strategy gradient, and game-enhanced control strategy generation to obtain a charging pile intelligent multiple control instruction set; the charging pile defense network diagram output module determines the PBFT consensus protocol through the charging pile intelligent multiple control instruction set, performs dynamic ant colony optimization, constructs a self-organizing defense network, and outputs a charging pile defense network diagram.

[0006] As a further solution, collect the multi-modal data of the charging pile intelligent monitoring and output the original charging pile intelligent monitoring status set. The specific analysis process includes: collecting the multi-modal data of the charging pile intelligent monitoring. The multi-modal data of the charging pile intelligent monitoring specifically includes the electrical data of the charging pile intelligent monitoring, the mechanical data of the charging pile intelligent monitoring, the environmental data of the charging pile intelligent monitoring, and the battery data of the charging pile intelligent monitoring. The electrical data of the charging pile intelligent monitoring specifically includes the charging pile monitoring voltage U jc , the charging pile monitoring current I jc , and the charging pile monitoring power factor η jc ; the mechanical data of the charging pile intelligent monitoring specifically includes the temperature Ts of the charging pile radiator, the corrosion area A of the charging pile shell fs , and the rotation speed v of the charging pile cooling fan s ; the environmental data of the charging pile intelligent monitoring specifically includes the humidity RH of the charging pile environment, the temperature H of the charging pile environment t , and the wind speed v of the charging pile environment f ; the battery data of the charging pile intelligent monitoring specifically includes the charging pile monitoring charging pile battery voltage SOC v , the current remaining power SOC of the charging pile battery d , and the temperature SOC of the charging pile battery T ; output the multi-modal data of the charging pile intelligent monitoring as the original charging pile intelligent monitoring status set.

[0007] As a further solution, based on the original charging pile intelligent monitoring status set, construct a charging pile intelligent monitoring risk matrix. The specific analysis process is as follows: based on the electrical data of the charging pile intelligent monitoring, comprehensively analyze to obtain the electrical risk factor f of the charging pile intelligent monitoring ele ; based on the mechanical data of the charging pile intelligent monitoring, comprehensively analyze to obtain the mechanical risk factor f of the charging pile intelligent monitoring Mac ; based on the environmental data of the charging pile intelligent monitoring, comprehensively analyze to obtain the environmental risk factor f of the charging pile intelligent monitoring env ; based on the battery data of the charging pile intelligent monitoring, comprehensively analyze to obtain the battery risk factor f of the charging pile intelligent monitoring SOC ; based on the electrical risk factor of the charging pile intelligent monitoring, the mechanical risk factor of the charging pile intelligent monitoring, the environmental risk factor of the charging pile intelligent monitoring, and the battery risk factor of the charging pile intelligent monitoring, construct a charging pile intelligent monitoring risk matrix M risk , M risk = [f ele f Mac f env f SOC .

[0008] As a further solution, the specific analysis process of the electrical risk factor of the charging pile intelligent monitoring is as follows:

[0009] Where U0 is the rated voltage of the charging pile stored in the database, I0 is the rated current of the charging pile stored in the database, and e is the natural constant.

[0010] As a further solution, the charging pile intelligently monitors the mechanical risk factors. The specific analysis process is as follows:

[0011] Where e is the natural constant and Sigmoid is the Sigmoid function.

[0012] As a further solution, the charging pile intelligently monitors the environmental risk factors. The specific analysis process is as follows:

[0013] Where e is the natural constant.

[0014] As a further solution, the charging pile intelligently monitors the battery risk factors. The specific analysis process is as follows:

[0015]

[0016] As a further solution, based on the intelligent monitoring risk matrix of the charging pile and obtaining the grid power, the deep deterministic policy gradient, the specific analysis process is as follows: Based on the intelligent monitoring risk matrix of the charging pile and obtaining the grid power P grid , perform the reconstruction of the charging pile state space s at time step t t : s t = [f ele , f Mac , f env , f SOC , P grid ; Quantize the charging pile control action a at time step t t , which specifically includes the charging power of the charging pile, the cooling intensity of the charging pile, and the electromagnetic shielding level of the charging pile;

[0017] Deep deterministic policy gradient: Policy network Actor:

[0018] a t = μ(s t | θ μ ) + ò t ; Where μ(·) is the neural network policy function, θ μ is the weight parameter of the policy network, and ò t is the Ornstein-Uhlenbeck process noise;

[0019] Value network Critic:

[0020] Q(s t , a t | θ Q ) = E[r t+γQ′(s t+1 ,μ′(s t+1 ))];

[0021] r t = w1P profit - w2(f ele + f Mac )); where Q(s t , a t |θ Q ) is the output of the value network, representing the value estimate when taking the charging pile control action a t at time step t in the charging pile state space s t at time step t, θ Q is the weight parameter of the value network, γ is the future reward discount factor, r t is the immediate reward function, w1 is the charging revenue weight, w2 is the risk penalty weight, P profit is the charging revenue per unit charging power, f ele is the intelligent monitoring electrical risk factor of the charging pile, f Mac is the intelligent monitoring mechanical risk factor of the charging pile, E is the expectation operator, Q′ is the output of the target value network, μ′ is the output of the target policy network, s t+1 is the charging pile state space at time step t + 1.

[0022] As a further solution, a game reinforcement control strategy is generated to obtain an intelligent multi-control instruction set for the charging pile. The specific analysis process is as follows: Establish a three-party non-cooperative game model:

[0023] Charging pile utility:

[0024] ; where U1 is the utility function value of the charging pile, k p is the power revenue coefficient, P set is the charging power set by the charging pile, k e is the electrical risk sensitivity, k m is the mechanical risk sensitivity, k v is the environmental risk sensitivity, f env is the intelligent monitoring environmental risk factor of the charging pile, and e is the natural constant

[0025] Grid utility:

[0026] where U2 is the utility function value of the grid, λ is the grid power balance coefficient, is the power that the grid can provide based on the grid power P grid collected;

[0027] User utility:

[0028] U3 = -f SOC -βP set ; where U3 is the user's utility function value, f SOC is the intelligent monitoring battery risk factor of the charging pile, and β is the user's electricity bill sensitivity coefficient;

[0029] Nash equilibrium solution:

[0030] where is the gradient of the strategy a i , U i is the utility function of the i-th party, i ∈ {1, 2, 3}, 1 corresponds to the charging pile, 2 corresponds to the power grid, 3 corresponds to the user, a i is the strategy of the i-th party, is the strategy combination of the other two parties except the i-th party;

[0031] H∞ robust controller control law:

[0032] where u(t) is the control input at time t, K is the optimal gain matrix obtained through the Riccati equation, and P real is the actual charging power of the charging pile;

[0033] Output the intelligent multiple control instruction set of the charging pile, specifically including the charging power P set ' of the charging pile, the cooling intensity T cool of the charging pile, and the electromagnetic shielding level Shield level of the charging pile.

[0034] As a further solution, through the intelligent multiple control instruction set of the charging pile, determine the PBFT consensus protocol, and perform dynamic ant colony optimization to construct a self-organizing defense network, and output the defense network diagram of the charging pile. The specific analysis process is as follows: PBFT consensus protocol:

[0035] Node verification condition:

[0036] N honest ≥ 2f + 1; where N honest is the number of honest nodes, and f is the maximum number of fault-tolerant malicious nodes;

[0037] Communication complexity:

[0038] C comm = O(N 2 );

[0039] N = 4f + 1; where C comm is the communication complexity of the PBFT consensus protocol, O(N 2 ) is the big O notation, and N is the total number of the minimum nodes;

[0040] Dynamic Ant Colony Optimization:

[0041] Pheromone Update:

[0042]

[0043] Where τ xy (t) is the pheromone concentration on the path from node x to node y at time t, and τ xy (t + 1) is the pheromone concentration on the path from node x to node y at time t + 1, ρ is the pheromone evaporation rate, is the pheromone increment left by the d-th ant on the path from node x to node y, and L xy is the physical distance from node x to node y, is the intelligent monitoring environmental risk factor of the charging pile at node y, and Shield level is the electromagnetic shielding level of the charging pile, and m is the total number of ants in the ant colony stored in the database, that is, the maximum number of iterations of the dynamic ant colony optimization;

[0044] Path Selection Probability:

[0045]

[0046]

[0047] Where is the probability that the d-th ant selects the path from node x to node y, α is the pheromone heuristic factor, β is the risk aversion factor, and η xy is the risk heuristic function from node x to node y, N d is the set of neighboring nodes related to the d-th ant, l belongs to N d , and τ xy is the pheromone concentration on the path from node x to node y, and τ xl is the pheromone concentration on the path from node x to node l, and η xl is the risk heuristic function from node x to node l, is the intelligent monitoring environmental risk factor of the charging pile at node l;

[0048] When the last iteration of the dynamic ant colony optimization is completed, the probability p xy of selecting the path from node x to node y is obtained; when p xy is greater than the path selection probability threshold stored in the database, then an edge between node x and node y is established, otherwise an edge between node x and node y is not established; at the same time, the weights of nodes x and y in the charging pile defense network graph are determined:

[0049] W xy = -ln(p xy ); Where Wxy is the weight from node x to node y;

[0050] Output the charging pile defense network graph G def , G def =(Vt, Et, Wt); where Vt is the set of all charging pile nodes, Et is the edge set of the charging pile defense network, and Wt is the weight matrix of the charging pile defense network; the network layer preferably selects W xy The minimum path communication transmission charging pile intelligent multiple control instruction set, and control the charging pile based on the charging pile intelligent multiple control instruction set.

[0051] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0052] (1) By providing a charging pile intelligent control system with multiple safety protections, the charging pile multimodal data acquisition module can collect multimodal data, covering various aspects of information such as electricity, machinery, and environment. Compared with a single data acquisition method, it can more comprehensively and accurately reflect the operating state of the charging pile. The monitoring risk matrix construction module constructs a risk matrix based on the collected original state set, which can systematically and intuitively present various risks faced by the charging pile and their degrees. The game reinforcement control strategy generation module combines information such as the risk matrix and grid power, and uses deep deterministic policy gradients to generate a control instruction set. It can be dynamically adjusted according to real-time changes, adapt to complex and changeable operating scenarios, ensure the stable and efficient operation of the charging pile, and improve the friendliness to the power grid.

[0053] (2) The charging pile defense network graph output module of the present invention constructs a self-organizing defense network by determining the PBFT consensus protocol and dynamic ant colony optimization. The PBFT consensus protocol can ensure that each node in the network can still reach a reliable consensus in the presence of some malicious or faulty nodes, guaranteeing the accuracy of system information interaction and decision-making. The dynamic ant colony optimization can adaptively adjust the defense strategy according to the risk situation and environmental changes, optimize resource allocation, enhance the ability of the charging pile system to resist external attacks and internal faults, and the output charging pile defense network graph visually displays the defense layout, which is convenient for management and maintenance. Description of the Drawings

[0054] The present invention is further described with the accompanying drawings, but the embodiments in the drawings do not constitute any limitation to the present invention. For those of ordinary skill in the art, without creative work, other drawings can also be obtained according to the following drawings.

[0055] Figure 1 is the system module connection schematic diagram of the present invention. Detailed Embodiments

[0056] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0057] Please refer to Figure 1 , the embodiments of the present invention provide an intelligent control technical solution for a charging pile with multiple safety protections: an intelligent control system for a charging pile with multiple safety protections, including a multi-modal data acquisition module for the charging pile, a monitoring risk matrix construction module, a game reinforcement control strategy generation module, and a charging pile defense network diagram output module.

[0058] The multi-modal data acquisition module for the charging pile collects multi-modal data for intelligent monitoring of the charging pile and outputs an original set of intelligent monitoring states of the charging pile.

[0059] Specifically, collecting multi-modal data for intelligent monitoring of the charging pile and outputting an original set of intelligent monitoring states of the charging pile. The specific analysis process includes: collecting multi-modal data for intelligent monitoring of the charging pile, and the multi-modal data for intelligent monitoring of the charging pile specifically includes electrical data for intelligent monitoring of the charging pile, mechanical data for intelligent monitoring of the charging pile, environmental data for intelligent monitoring of the charging pile, and battery data for intelligent monitoring of the charging pile; the electrical data for intelligent monitoring of the charging pile specifically includes the charging pile monitoring voltage U jc , the charging pile monitoring current I jc , the charging pile monitoring power factor η jc ; the mechanical data for intelligent monitoring of the charging pile specifically includes the temperature Ts of the charging pile radiator, the corrosion area A of the charging pile housing fs , the rotation speed v of the charging pile cooling fan s ; the environmental data for intelligent monitoring of the charging pile specifically includes the humidity RH of the charging pile environment, the temperature H of the charging pile environment t , the wind speed v of the charging pile environment f ; the battery data for intelligent monitoring of the charging pile specifically includes the charging pile monitoring battery voltage SOC v , the current remaining power of the charging pile battery SOC d , the temperature of the charging pile battery SOC T ; output the multi-modal data for intelligent monitoring of the charging pile as an original set of intelligent monitoring states of the charging pile.

[0060] The data collected covers electrical, mechanical, environmental, and battery data. Electrical data include voltage, current, and power factor, which can accurately reflect the power transmission and usage status of the charging pile; mechanical data include radiator temperature, shell corrosion area, and cooling fan speed, which can monitor the operation of the mechanical parts of the equipment; environmental data include humidity, temperature, and wind speed, which can understand the impact of the external environment on the charging pile; battery data includes voltage, remaining power, and temperature, which can grasp the battery health and charging and discharging status. All-round data collection avoids the omission of key information and fully outlines the overall operation of the charging pile.

[0061] By continuously monitoring multimodal data, subtle changes in parameters can be captured. For example, abnormal fluctuations in electrical data may indicate line failure or equipment aging; deviations from the normal range of mechanical parameters, such as excessive radiator temperature and abnormal fan speed, may indicate a failure in the cooling system; extreme changes in environmental parameters will affect the performance of charging piles; abnormal battery data reflects battery health issues. Based on these data changes, faults can be warned in advance, facilitating timely troubleshooting and maintenance, and reducing the occurrence rate and degree of harm of faults.

[0062] Provide real-time data for the intelligent control system of the charging pile, and the system adjusts the operating parameters based on this data. Adjust heat dissipation and protection measures according to the ambient temperature and humidity; optimize the charging strategy according to the battery status to improve charging efficiency and battery life. Realize intelligent and adaptive operation of the charging pile and improve overall performance.

[0063] The monitoring risk matrix construction module constructs the charging pile intelligent monitoring risk matrix based on the original charging pile intelligent monitoring status set.

[0064] Specifically, based on the original charging pile intelligent monitoring state set, the charging pile intelligent monitoring risk matrix is constructed. The specific analysis process is as follows: Based on the charging pile intelligent monitoring electrical data, the charging pile intelligent monitoring electrical risk factor f is obtained through comprehensive analysis ele Based on the data of intelligent monitoring machinery of charging piles, the risk factor f of intelligent monitoring machinery of charging piles is obtained through comprehensive analysis Mac Based on the intelligent monitoring environment data of charging piles, a comprehensive analysis is conducted to obtain the intelligent monitoring environment risk factor f of charging piles env Based on the charging pile intelligent monitoring battery data, a comprehensive analysis is performed to obtain the charging pile intelligent monitoring battery risk factor f SOC Based on the charging pile intelligent monitoring electrical risk factor, charging pile intelligent monitoring mechanical risk factor, charging pile intelligent monitoring environmental risk factor, and charging pile intelligent monitoring battery risk factor, the charging pile intelligent monitoring risk matrix M is constructed. risk , M risk =[f ele f Mac f env f SOC ], as the analytical basis for obtaining the intelligent multiple control instruction set of the charging pile.

[0065] Furthermore, the charging pile intelligently monitors the electrical risk factors, and the specific analysis process is as follows:

[0066] In the formula, U0 is the rated voltage of the charging pile stored in the database, I0 is the rated current of the charging pile stored in the database, and e is the natural constant.

[0067] Specifically, the charging pile intelligently monitors the mechanical risk factors, and the specific analysis process is as follows:

[0068] In the formula, e is the natural constant, and Sigmoid is the Sigmoid function.

[0069] Specifically, the charging pile intelligently monitors the environmental risk factors, and the specific analysis process is as follows:

[0070] In the formula, e is the natural constant.

[0071] Furthermore, the charging pile intelligently monitors the battery risk factors, and the specific analysis process is as follows:

[0072]

[0073] By calculating the risk factors in different dimensions such as electrical, mechanical, environmental, and battery, various risks faced by the charging pile are transformed from qualitative descriptions into quantitative values. For example, the electrical risk factor is calculated by combining the voltage, current, and deviation from the rated value, which can accurately reflect the abnormal degree of the electrical system; the mechanical risk factor comprehensively evaluates the risk of mechanical components by considering the radiator temperature, the corrosion area of the shell, etc., so that the risk level can be measured by specific numerical values, which is convenient for accurately grasping the risk size.

[0074] Based on the multi-dimensional risk factors, a matrix is constructed for comprehensive analysis. Considering the mutual correlation and influence among different risks, rather than looking at each risk in isolation. Electrical faults may cause mechanical problems, and environmental factors will also affect the electrical and mechanical performance. Comprehensive evaluation can formulate more reasonable and effective risk response strategies. The risk matrix provides a key analysis basis for generating the intelligent multiple control instruction set of the charging pile. The system adjusts the operating parameters of the charging pile and takes protective measures according to the risk status reflected by the matrix.

[0075] The game reinforcement control strategy generation module, based on the intelligent monitoring risk matrix of the charging pile, obtains the grid power, and through deep deterministic policy gradient, generates the game reinforcement control strategy to obtain the intelligent multiple control instruction set of the charging pile.

[0076] Specifically, based on the intelligent monitoring risk matrix of the charging pile, obtaining the grid power, and the deep deterministic policy gradient, the specific analysis process is as follows: Based on the intelligent monitoring risk matrix of the charging pile, and obtaining the grid power P grid, perform the charging pile state space s at time step t t Reconstruction: s t =[f ele , f Mac , f env , f SOC , P grid ; the charging pile control action a at time step t t Quantization, specifically including the charging power of the charging pile, the cooling intensity of the charging pile, and the electromagnetic shielding level of the charging pile;

[0077] Deep Deterministic Policy Gradient: Policy Network Actor:

[0078] a t =μ(s t |θ μ ) + ò t ; In the formula, μ(·) is the neural network policy function, θ μ is the weight parameter of the policy network, ò t is the Ornstein-Uhlenbeck process noise;

[0079] Value Network Critic:

[0080] Q(s t , a t |θ Q ) = E[r t +γQ′(s t+1 , μ′(s t+1 ))];

[0081] r t = w1P profit - w2(f ele + f Mac ); In the formula, Q(s t , a t |θ Q ) is the output of the value network, representing the value estimation when taking the charging pile control action a t at time step t in the charging pile state space s t at time step t, θ Q is the value network weight parameter, γ is the future reward discount factor, r t is the immediate reward function, w1 is the charging revenue weight, w2 is the risk penalty weight, P profit is the charging revenue per unit charging power, f ele is the intelligent monitoring electrical risk factor of the charging pile, f Mac is the intelligent monitoring mechanical risk factor of the charging pile, E is the expectation operator, Q′ is the output of the target value network, μ′ is the output of the target policy network, s t+1 is the charging pile state space at time step t + 1.

[0082] Analyze by combining the intelligent monitoring risk matrix of the charging pile and the power grid power information. Not only focus on the risks in aspects such as the charging pile's own electricity, machinery, environment, battery, etc., but also consider the power that the power grid can provide. It can find a balance between the risk state of the charging pile and the operating condition of the power grid, avoid affecting the stability of the power grid due to the operation of the charging pile, or the charging pile being unable to work properly due to the power grid power limit, and achieve system-level optimal decision-making.

[0083] Apply the Deep Deterministic Policy Gradient (DDPG) algorithm to continuously optimize and adapt the control actions of the charging pile to complex and changeable operating scenarios. For example, in the face of different risk combinations and power grid power fluctuations, it can dynamically adjust control actions such as charging power, cooling intensity, and electromagnetic shielding level to achieve intelligent adaptive control. Evaluate the action value output by the policy network. Consider the immediate reward and the future reward discount value, comprehensively balance the charging income and the risk penalty, and balance the income and the risk. This evaluation mechanism guides the policy network to learn better strategies and improve the comprehensive benefit of the charging pile operation.

[0084] Furthermore, generate a game reinforcement control strategy to obtain an intelligent multiple control instruction set for the charging pile. The specific analysis process is as follows: Establish a three-party non-cooperative game model:

[0085] Charging pile utility:

[0086] ; where U1 is the utility function value of the charging pile, k p is the power income coefficient, P set is the charging power set by the charging pile, k e is the electrical risk sensitivity, k m is the mechanical risk sensitivity, k v is the environmental risk sensitivity, f env is the intelligent monitoring environmental risk factor of the charging pile, and e is the natural constant

[0087] Power grid utility:

[0088] In the formula, U2 is the utility function value of the power grid, λ is the power grid power balance coefficient, is the power that the power grid can provide based on the power grid power P grid collected;

[0089] User utility:

[0090] U3 = -f SOC - βP set ; where U3 is the utility function value of the user, f SOC is the intelligent monitoring battery risk factor of the charging pile, and β is the user's electricity bill sensitivity coefficient;

[0091] Nash equilibrium solution:

[0092] In the formula, is the gradient of the strategy a i seeking, U i is the utility function of the i-th party, i ∈ {1, 2, 3}, 1 corresponds to the charging pile, 2 corresponds to the power grid, 3 corresponds to the user, a i is the strategy of the i-th party, is the strategy combination of the other two parties except the i-th party;

[0093] H∞ robust controller control law:

[0094] In the formula, u(t) is the control input at time t, K is the optimal gain matrix obtained through the Riccati equation, P real is the actual charging power of the charging pile;

[0095] Output the intelligent multi-control instruction set of the charging pile, specifically including the charging power P set ′ of the charging pile, the cooling intensity T cool of the charging pile, and the electromagnetic shielding level Shield level of the charging pile.

[0096] By establishing a three-party non-cooperative game model, the utility functions of the charging pile, the power grid, and the user are constructed respectively. The utility of the charging pile comprehensively considers the power income and the penalties for electrical, mechanical, and environmental risks; the utility of the power grid focuses on power balance; the utility of the user involves battery risks and electricity bills. Comprehensively considering the interests of the three parties, avoiding excessive damage to other parties by one party, and promoting the overall coordinated development of the system.

[0097] Using the Nash equilibrium solution to find the stable equilibrium point of the strategies of the three parties. At this point, any party unilaterally changing its strategy cannot increase its own utility, ensuring the relative balance of the interests of the three parties, and making the operation of the charging pile, the power supply of the power grid, and the use of the user in a relatively stable and reasonable state, improving the stability and sustainability of the system.

[0098] Adopting the H∞ robust controller control law to generate the control input based on information such as risk factors and power deviation. It can effectively cope with the uncertainties and disturbances in the system, such as electrical parameter fluctuations, mechanical fault hazards, environmental changes, etc. Adjusting the control strategy through the optimal gain matrix K, enabling the charging pile to still operate stably under complex and changeable working conditions, reducing the impact of risks, and improving the anti-interference ability and robustness of the system.

[0099] The final output includes an intelligent multi - control instruction set for charging piles, such as charging power, cooling intensity, electromagnetic shielding level, etc. According to the results of the three - party game equilibrium and robust control strategies, the operating parameters of the charging piles are precisely adjusted to achieve intelligent and refined control. Improve the operating efficiency and safety of charging piles, such as reasonably adjusting the charging power to avoid grid overload, regulating the cooling intensity to ensure the normal temperature of the equipment, and enhancing the electromagnetic shielding level to reduce electromagnetic interference.

[0100] The charging pile defense network graph output module determines the PBFT consensus protocol through the intelligent multi - control instruction set of the charging pile, and performs dynamic ant colony optimization to construct a self - organizing defense network and output the charging pile defense network graph.

[0101] Specifically, through the intelligent multi - control instruction set of the charging pile, the PBFT consensus protocol is determined, dynamic ant colony optimization is performed, a self - organizing defense network is constructed, and the charging pile defense network graph is output. The specific analysis process is as follows: PBFT consensus protocol:

[0102] Node verification condition:

[0103] N honest ≥2f + 1; where N honest is the number of honest nodes, and f is the maximum number of fault - tolerant malicious nodes stored in the database;

[0104] Communication complexity:

[0105] C comm =O(N 2 );

[0106] N = 4f + 1; where C comm is the communication complexity of the PBFT consensus protocol, O(N 2 ) is the big O notation, and N is the total number of the smallest nodes;

[0107] Dynamic ant colony optimization:

[0108] Pheromone update:

[0109]

[0110] where τ xy (t) is the pheromone concentration on the path from node x to node y at time t, τ xy (t + 1) is the pheromone concentration on the path from node x to node y at time t + 1, ρ is the pheromone evaporation rate, is the pheromone increment left by the d - th ant on the path from node x to node y, L xy is the physical distance from node x to node y, is the intelligent monitoring environmental risk factor of the charging pile at node y, Shield levelThe electromagnetic shielding level of the charging pile, m is the total number of ants in the ant colony stored in the database, that is, the maximum number of iterations of the dynamic ant colony optimization;

[0111] Path selection probability:

[0112]

[0113] In the formula, is the probability that the d-th ant selects the path from node x to node y. α is the pheromone heuristic factor, β is the risk aversion factor, and η xy is the risk heuristic function from node x to node y. N d is the set of neighborhood nodes related to the d-th ant. l belongs to N d , τ xy is the pheromone concentration on the path from node x to node y. τ xl is the pheromone concentration on the path from node x to node l. η xl is the risk heuristic function from node x to node l. is the risk factor of the intelligent monitoring environment of the charging pile at node l;

[0114] When the last iteration of the dynamic ant colony optimization is completed, the probability p of selecting the path from node x to node y is obtained xy ; when p xy is greater than the path selection probability threshold stored in the database, then the edge between node x and node y is established, otherwise the edge between node x and node y is not established; at the same time, determine the weights of nodes x and y in the charging pile defense network graph:

[0115] W xy = -ln(p xy ); in the formula, W xy is the weight from node x to node y;

[0116] Output the charging pile defense network graph G def G def = (Vt, Et, Wt); where, Vt is the set of all charging pile nodes, Et is the edge set of the charging pile defense network, and Wt is the weight matrix of the charging pile defense network; the network layer preferentially selects the path with the smallest W xy to communicate and transmit the intelligent multiple control instruction set of the charging pile, and control the charging pile based on the intelligent multiple control instruction set of the charging pile.

[0117] Through the PBFT consensus protocol, meeting the node verification conditions, it can ensure that nodes in the network reach a reliable consensus in the presence of a certain number of malicious nodes. Prevent malicious nodes from interfering with the information transmission and decision-making process, ensure the security and stability of the charging pile system at the network level, and reduce the risk of system paralysis or incorrect operation caused by node failures or malicious behaviors.

[0118] The dynamic ant colony optimization mechanism calculates based on pheromone update and path selection probability, comprehensively considering factors such as physical distance, environmental risk factor, and electromagnetic shielding level. It can adaptively adjust the defense strategy according to the actual operating environment and risk status of the charging pile, guide information to be transmitted along paths with lower risks and better distances, effectively avoid high-risk paths and nodes, and improve the system's ability to respond to environmental risks.

[0119] The constructed self-organizing defense network clarifies the connection relationship and weights between nodes. The network layer preferentially selects the path with the smallest weight for communication to transmit control instructions. It realizes the efficient transmission of control instructions, enables the intelligent multiple control instructions of the charging pile to be issued promptly and accurately, and can also allocate resources such as communication and calculation according to the quality of the path, improve resource utilization efficiency, and ensure the efficient operation of the charging pile system.

[0120] The module automatically constructs a self-organizing defense network without excessive manual intervention, and can dynamically adjust the network structure and parameters according to real-time risks and environmental changes. For example, it establishes or disconnects the connections between nodes based on the results of dynamic ant colony optimization, determines the node weights, enables the system to have the ability of self-adaptation and optimization, and improves the degree of intelligence.

[0121] The above content is only an example and explanation of the structure of the present invention. Those skilled in the art of this technology can make various modifications or supplements to the described specific embodiments or use similar methods to replace them. As long as they do not deviate from the structure of the invention or exceed the scope defined by this claim book, they shall fall within the protection scope of the present invention.

Claims

1. An intelligent control system for a charging pile with multiple safety protections, characterized in that, It includes a charging pile multi-modal data acquisition module, a monitoring risk matrix construction module, a game reinforcement control strategy generation module, and a charging pile defense network diagram output module: The charging pile multi-modal data acquisition module collects multi-modal data of intelligent monitoring of charging piles and outputs the original intelligent monitoring state set of charging piles; The monitoring risk matrix construction module constructs a risk matrix for intelligent monitoring of charging piles based on the original intelligent monitoring state set of charging piles; The game reinforcement control strategy generation module, based on the risk matrix of intelligent monitoring of charging piles and obtaining the grid power, performs deep deterministic policy gradient and game reinforcement control strategy generation to obtain a set of intelligent multiple control instructions for charging piles; The charging pile defense network diagram output module determines the PBFT consensus protocol through the set of intelligent multiple control instructions for charging piles, performs dynamic ant colony optimization, constructs a self-organizing defense network, and outputs a charging pile defense network diagram.

2. The intelligent control system of a charging pile with multiple safety protections according to claim 1, wherein: The process of collecting multi-modal data of intelligent monitoring of charging piles and outputting the original intelligent monitoring state set of charging piles is specifically analyzed as follows: Collect multi-modal data of intelligent monitoring of charging piles. The multi-modal data of intelligent monitoring of charging piles specifically includes electrical data of intelligent monitoring of charging piles, mechanical data of intelligent monitoring of charging piles, environmental data of intelligent monitoring of charging piles, and battery data of intelligent monitoring of charging piles; Charging pile intelligent monitoring electrical data specifically includes charging pile monitoring voltage U jc , Charging pile monitoring current I jc , Charging pile monitoring power factor η jc ; The specific mechanical data for intelligent monitoring of the charging pile includes the temperature Ts of the charging pile radiator and the corrosion area A of the charging pile housing fs , the rotation speed v of the charging pile cooling fan s ; The specific environmental data monitored by the charging pile intelligent monitoring includes the environmental humidity RH of the charging pile and the environmental temperature H of the charging pile t and the environmental wind speed v of the charging pile f ; The intelligent monitoring of the battery data by the charging pile specifically includes the charging pile monitoring the battery voltage SOC of the charging pile v and the current remaining power SOC of the charging pile battery d and the temperature SOC of the charging pile battery T ; Output the multi-modal data of intelligent monitoring of charging piles as the original intelligent monitoring state set of charging piles.

3. The intelligent control system for a charging pile with multiple safety protections according to claim 2, characterized in that: The process of constructing a risk matrix for intelligent monitoring of charging piles based on the original intelligent monitoring state set of charging piles is specifically analyzed as follows: Based on the intelligent monitoring of electrical data of the charging pile, the intelligent monitoring electrical risk factor f of the charging pile is obtained through comprehensive analysis ele ; Based on the intelligent monitoring mechanical data of the charging pile, the intelligent monitoring mechanical risk factor f of the charging pile is obtained through comprehensive analysis Mac ; Based on the intelligent monitoring environmental data of the charging pile, the intelligent monitoring environmental risk factor f of the charging pile is obtained through comprehensive analysis env ; Based on the intelligent monitoring of battery data by the charging pile, the intelligent monitoring battery risk factor f of the charging pile is obtained through comprehensive analysis SOC ; Based on the intelligent monitoring electrical risk factors of charging piles, the intelligent monitoring mechanical risk factors of charging piles, the intelligent monitoring environmental risk factors of charging piles, and the intelligent monitoring battery risk factors of charging piles, construct the intelligent monitoring risk matrix M of charging piles risk , M risk = [f ele f Mac f env f SOC .

4. The intelligent control system for a charging pile with multiple safety protections according to claim 3, characterized in that: The process of specifically analyzing the electrical risk factor of intelligent monitoring of charging piles is as follows: In the formula, U0 is the rated voltage of the charging pile stored in the database, I0 is the rated current of the charging pile stored in the database, and e is the natural constant.

5. The intelligent control system of a charging pile with multiple safety protections according to claim 3, characterized in that: The process of specifically analyzing the mechanical risk factor of intelligent monitoring of charging piles is as follows: In the formula, e is the natural constant and Sigmoid is the Sigmoid function.

6. The intelligent control system of a charging pile with multiple safety protections according to claim 3, characterized in that: The process of specifically analyzing the environmental risk factor of intelligent monitoring of charging piles is as follows: In the formula, e is the natural constant.

7. An intelligent control system for a charging pile with multiple safety protections according to claim 3, characterized in that: The process of specifically analyzing the battery risk factor of intelligent monitoring of charging piles is as follows:

8. The intelligent control system for a charging pile with multiple safety protections according to claim 1, wherein: The process of specifically analyzing based on the risk matrix of intelligent monitoring of charging piles and obtaining the grid power and performing deep deterministic policy gradient is as follows: Based on the intelligent monitoring risk matrix of the charging pile and obtain the grid power P grid , perform the charging pile state space s at time step t t Reconstruction: s t = [f ele , f Mac , f env , f SOC , P grid ; Charging pile control action a at time step t t Quantification, specifically including the charging power of the charging pile, the cooling intensity of the charging pile, and the electromagnetic shielding level of the charging pile; Deep deterministic policy gradient: Policy network Actor: a t = μ(s t | θ μ ) + ò t ; where μ(·) is the neural network policy function, and θ μ are the weight parameters of the policy network, and ò t is the Ornstein-Uhlenbeck process noise; Value network Critic: Q(s t , a t | θ Q ) = E[r t + γQ'(s t+1 , μ'(s t+1 ))]; r t = w1P profit - w2(f ele + f Mac ); Where, Q(s t , a t |θ Q ) is the output of the value network, representing the value estimate when taking the charging pile control action a t at time step t under the charging pile state space s t at time step t, θ Q is the weight parameter of the value network, γ is the future reward discount factor, r t is the immediate reward function, w1 is the charging revenue weight, w2 is the risk penalty weight, P profit is the charging revenue per unit charging power, f ele is the electrical risk factor for intelligent monitoring of the charging pile, f Mac is the mechanical risk factor for intelligent monitoring of the charging pile, E is the expectation operator, Q' is the output of the target value network, μ' is the output of the target policy network, s t+1 is the charging pile state space at time step t + 1.

9. The intelligent control system for a charging pile with multiple safety protections according to claim 8, wherein: The process of generating the game reinforcement control strategy to obtain a set of intelligent multiple control instructions for charging piles is specifically analyzed as follows: Establish a three-party non-cooperative game model: Charging pile utility: Where, U1 is the utility function value of the charging pile, and k p is the power revenue coefficient, P set is the charging power set by the charging pile, k e is the electrical risk sensitivity, k m is the mechanical risk sensitivity, k v is the environmental risk sensitivity, f env is the environmental risk factor for intelligent monitoring of the charging pile, and e is the natural constant Grid utility: Wherein, U2 is the utility function value of the power grid, and λ is the power grid power balance coefficient. is the available power provided by the power grid based on the power grid power P grid collected. User utility: U3 = -f SOC -βP set ; where U3 is the user's utility function value, f SOC is the intelligent monitoring battery risk factor of the charging pile, and β is the user's electricity charge sensitivity coefficient; Nash equilibrium solution: In the formula, is to find the gradient of policy a i U i is the utility function of the i-th party, where i ∈ {1, 2, 3}, 1 corresponds to the charging pile, 2 corresponds to the power grid, 3 corresponds to the user, and a i is the policy of the i-th party, is the policy combination of the other two parties except the i-th party; H∞ robust controller control law: where \(u(t)\) is the control input at time \(t\), \(K\) is the optimal gain matrix obtained by the Riccati equation, and \(P\) real is the actual charging power of the charging pile; Output intelligent multiple control instruction sets for charging piles, specifically including the charging power P of the charging pile set ′, the cooling intensity T of the charging pile cool , the electromagnetic shielding level Shield of the charging pile level .

10. The intelligent control system for a charging pile with multiple safety protections according to claim 9, wherein: The process of determining the PBFT consensus protocol through the set of intelligent multiple control instructions for charging piles, performing dynamic ant colony optimization, constructing a self-organizing defense network, and outputting a charging pile defense network diagram is specifically analyzed as follows: PBFT consensus protocol: Node verification conditions: N honest ≥2f + 1; where N honest is the number of honest nodes, and f is the maximum number of malicious nodes that can be tolerated; Communication complexity: C comm = O(N 2 ); N = 4f + 1; Where C comm is the communication complexity of the PBFT consensus protocol, O(N 2 ) is the big O notation, and N is the total number of minimum nodes; Dynamic ant colony optimization: Pheromone update: where τ xy (t) is the pheromone concentration on the path from node x to node y at time t, and τ xy (t + 1) is the pheromone concentration on the path from node x to node y at time t + 1. ρ is the pheromone evaporation rate, is the pheromone increment left by the d-th ant on the path from node x to node y. L xy is the physical distance from node x to node y, is the intelligent monitoring environmental risk factor of the charging pile at node y. Shield level is the electromagnetic shielding level of the charging pile. m is the total number of ants in the ant colony stored in the database, that is, the maximum number of iterations of the dynamic ant colony optimization; Path selection probability: Wherein, is the probability that the d-th ant selects the path from node x to node y, α is the pheromone heuristic factor, β is the risk aversion factor, η xy is the risk heuristic function from node x to node y, N d is the set of neighborhood nodes related to the d-th ant, l belongs to N d , τ xy is the pheromone concentration on the path from node x to node y, τ xl is the pheromone concentration on the path from node x to node l, η xl is the risk heuristic function from node x to node l, is the intelligent monitoring environmental risk factor of the charging pile at node l; When the last iteration of the dynamic ant colony optimization is completed, the probability p of selecting the path from node x to node y is obtained xy ; When p xy is greater than the path selection probability threshold stored in the database, an edge between node x and node y is established; otherwise, no edge between node x and node y is established. Simultaneously determine the weights of node x and node y of the charging pile defense network diagram: W xy = -ln(p xy ); where W xy is the weight from node x to node y; Output charging pile defense network graph G def , G def =(Vt, Et, Wt); Among them, Vt is the set of all charging pile nodes, Et is the edge set of the charging pile defense network, and Wt is the weight matrix of the charging pile defense network; The network layer preferentially selects W xy The smallest path communication transmission charging pile intelligent multiple control instruction set is used to control the charging pile based on the charging pile intelligent multiple control instruction set.

Images