A charging and discharging control system and electronic equipment

Abstract

The present invention discloses a charging and discharging control system and electronic device. S1. Constructing a multi-objective optimization model for a charging pile; S2. Initializing a chaotic particle swarm optimization algorithm; S3. Calculating each particle in the particle swarm according to the objective function in the multi-objective optimization model for the charging pile; S4. Iterating repeatedly until the chaotic particle swarm optimization algorithm reaches convergence conditions; S5. Generating a prediction result of the charging pile operating environment; S6. Updating the current optimal particle and the global optimal particle; S7. Generating a multi-objective charging and discharging optimization control strategy for the charging pile based on the optimized global optimal particle, applying the optimization control strategy to a charging pile charging and discharging control execution module to execute the multi-objective charging and discharging control of the charging pile. The present invention can quickly solve the global optimal solution in complex scenarios, meeting the requirements for efficient operation of electric vehicle charging piles in large-scale deployment scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of charging piles, and in particular to a charging and discharging control system and electronic equipment. Background Art

[0002] With the increasing penetration of electric vehicles and increasing pressure on power grids, optimal charging and discharging control technology for charging piles has become a key issue in smart grid management. Existing charging pile optimization control methods primarily focus on scheduling strategies based on single-objective optimization, such as minimizing energy consumption or charging time. These methods employ fixed control rules based on static models. While these methods can meet certain usage requirements in specific scenarios, their adaptability and optimization effectiveness are significantly limited in dynamic and complex power environments.

[0003] At present, some technologies attempt to introduce intelligent algorithms to improve optimization effects, including traditional particle swarm optimization algorithms and genetic algorithms. By iteratively searching for the global optimal solution, the scheduling efficiency of charging piles has been improved to a certain extent. However, traditional intelligent algorithms are prone to falling into local optimality in high-dimensional multi-objective optimization problems and lack the ability to adapt to dynamic constraints. In addition, existing technologies generally lack comprehensive consideration of user charging needs, grid load fluctuations and dynamic changes in electricity prices. The real-time performance and flexibility of the optimization model are poor, making it difficult to cope with changing charging scenarios.

[0004] Other technologies use prediction-based load balancing strategies, modeling and predicting grid load and user demand through historical data. Most of them are based on static prediction models with low utilization of real-time data and are difficult to dynamically adjust in uncertain environments. At the same time, existing technologies lack a multi-objective comprehensive optimization mechanism for user demand, grid load and electricity price fluctuations, making it impossible to effectively balance conflicts between various objectives in actual applications.

[0005] In summary, the existing technology mainly has the following shortcomings:

[0006] 1. Most optimization models are single-objective oriented and lack multi-objective comprehensive optimization capabilities, making it difficult to meet the multiple requirements of energy consumption, grid load, and user satisfaction.

[0007] 2. Intelligent algorithms lack global search capabilities under complex constraints and are prone to falling into local optimality, resulting in poor global and dynamic adaptability of optimization results.

[0008] 3. Limited dynamic prediction and real-time adjustment capabilities make it impossible to dynamically optimize control strategies based on real-time changes in grid load and user demand, resulting in low system robustness and responsiveness in complex environments.

[0009] 4. Lack of effective utilization of multi-source heterogeneous data, inability to fully tap the value of real-time data, and optimization effects are constrained by data incompleteness.

[0010] The above problems limit the further development of charging and discharging optimization control technology of charging piles. It is urgent to propose a multi-objective optimization control method that can adapt to dynamic and complex scenarios to solve the shortcomings of existing technologies. Summary of the Invention

[0011] One object of the present invention is to provide a charging and discharging control system and electronic equipment, which can quickly solve the global optimal solution in complex scenarios and meet the requirements for efficient operation of electric vehicle charging piles in large-scale deployment scenarios.

[0012] A charge and discharge control system according to an embodiment of the present invention includes the following steps:

[0013] S1. Construct a multi-objective optimization model for charging piles, parameterize the multi-objective optimization problem into a multi-objective optimization model for charging piles, and define the objective function and constraints.

[0014] S2. Initialize the chaotic particle swarm optimization algorithm, set the number of particles, initial particle positions, particle velocity range, and chaos mapping function, generate an initial particle distribution with global search capability based on the chaos mapping, and use the initial particle distribution to solve the multi-objective optimization model of the charging pile;

[0015] S3. Calculate the fitness value of each particle in the particle swarm based on the objective function in the multi-objective optimization model of the charging pile, sort the particles according to the fitness value, select the current optimal particle and the global optimal particle, and store them as the initial candidate solutions of the global optimal solution;

[0016] S4. Using the initial candidate solution as the reference solution, combined with the chaotic perturbation mechanism, the current position and velocity of the particle are updated through a nonlinear chaotic mapping function. The updated particle state is used for the next fitness value calculation. The fitness of the reference solution is then compared and screened to enhance the particle swarm's ability to escape from the local optimal solution. This process is repeated until the chaotic particle swarm optimization algorithm reaches convergence.

[0017] S5. Use the Bayesian inference model to model and predict the dynamic data in the charging pile operating environment in real time, and generate prediction results of the charging pile operating environment;

[0018] S6. Introduce the charging pile operating environment prediction results generated by the Bayesian inference model into the chaotic particle swarm optimization algorithm, dynamically adjust the objective weights and constraints in the charging pile multi-objective optimization model, use the updated optimization model to recalculate the particle fitness values, and update the current optimal particle and the global optimal particle;

[0019] S7. Generate a multi-objective charging and discharging optimization control strategy for the charging pile based on the optimized global optimal particles, apply the optimization control strategy to the charging pile charging and discharging control execution module, and execute the multi-objective charging and discharging control of the charging pile.

[0020] Optionally, the S1 includes:

[0021] S11. Construct a multi-objective optimization model for charging piles and parameterize the optimization objectives into objective functions, including the energy consumption minimization objective function minF1, the grid load balancing objective function minF2, and the user satisfaction maximization objective function minF3;

[0022] S12. Define the constraints of the multi-objective optimization model for charging piles, including:

[0023] Charging power constraints:

[0024] P min ≤P t ≤P max ;

[0025] Among them, P t is the charging power at time t, P min and P max are the minimum and maximum charging power allowed by the charging pile respectively;

[0026] Grid load limit:

[0027] L t ≤L max ;

[0028] Among them, L t is the grid load at time t, L max is the maximum load that the power grid can bear;

[0029] User charging requirements meet the following constraints:

[0030]

[0031] Among them, E i is the charging amount obtained by the i-th user during his charging time period, and are the charging start time and end time of the i-th user, The minimum charging amount required for the i-th user;

[0032] S13. Convert the multi-objective optimization problem into a multi-objective optimization model for charging piles, comprehensively consider the three objectives of minimizing energy consumption, balancing the grid load, and maximizing user satisfaction, and convert the multi-objective optimization problem into a single-objective optimization problem using the weighted summation method:

[0033] minF=w1minF1+w2minF2-w3minF3;

[0034] Among them, w1, w2, and w3 are target weight coefficients;

[0035] S14. Define the decision variables of the multi-objective optimization model of the charging pile as the charging power P in each time period. t .

[0036] Optionally, the S2 includes:

[0037] S21, initialize the chaotic particle swarm optimization algorithm and set the number of particles N p , the initial position of the particle X i (0) and initial velocity V i (0), where i represents the i-th particle, X i The initial value of (0) is randomly distributed in the feasible solution space, V i The initial value of (0) is within the preset speed range [V min ,V max ] Randomly generated within;

[0038] S22. Define the chaotic mapping function C(k) using the Logistic mapping form:

[0039] C(k+1)=μC(k)(1-C(k));

[0040] Where k is the number of iterations, C(k) represents the chaotic variable, with an initial value of C(0)∈(0,1)), and μ is the control parameter of the chaotic mapping, with a value range of (0,4];

[0041] S23. Generate a chaotic sequence according to the chaotic mapping function, and use the chaotic sequence to adjust the initial position and initial velocity of the particle. The particle position is updated as follows:

[0042] X i (0) = X min +(X max -X min )·C(k);

[0043] The particle velocity is updated as:

[0044] V i (0) = V min +(V max -V min )·C(k);

[0045] Among them, X min and X max are the upper and lower bounds of the position, V min and V max are the upper and lower bounds of the velocity respectively;

[0046] S24, using the initialized particle position X i (0) and initial velocity V i (0) The initial particle distribution is introduced into the objective function of the multi-objective optimization model of the charging pile, and the fitness value F(X i (0)) and sort the particles by fitness, and select the initial global optimal particle G(0) and the initial individual optimal particle P i (0);

[0047] S25. Taking the initialized chaotic particle swarm state as input, an initial particle swarm with global search capability is formed.

[0048] Optionally, the S3 includes:

[0049] S31. For each particle in the particle swarm, calculate the fitness value F(X i ):

[0050] F(X i )=w4F1(X i )+w5F2(X i )-w6F3(X i );

[0051] Among them, X i is the current position of the i-th particle, F1(X i )、F2(X i )、F3(X i ) are the energy consumption minimization objective function, the grid load balancing objective function and the user satisfaction maximization objective function at position X. i The value at , w4, w5, w6 are the target weight coefficients;

[0052] S32, based on the fitness value F(X i ) Sort the particles, the smaller the fitness value, the better the performance of the particle;

[0053] S33, record the individual historical optimal position P of each particle i And the corresponding optimal fitness value F(P i ), where P i is the position with the minimum fitness value found by the i-th particle so far;

[0054] S34. Select the particle with the smallest fitness value in the current particle swarm as the current global optimal particle G, and record the initial candidate solution X of the global optimal solution. gbest And the fitness value F gbest .

[0055] Optionally, the S4 includes:

[0056] S41, the initial candidate solution X of the global optimal solution gbest As a reference solution, the current position X of the particle is calculated by combining the chaotic perturbation mechanism. i (k) and speed V i (k) is updated, where k represents the current iteration number;

[0057] S42, generating a chaotic sequence C1(k) using a nonlinear chaotic mapping function;

[0058] S43. Update the particle velocity V based on the chaotic sequence C1(k) i (k+1) and position X i (k+1):

[0059] V i (k+1)=w·V i (k)+c1·r1·(P i -X i (k))+c2·r2·(X gbest -X i (k))+C1(k);

[0060] X i (k+1)=X i (k)+V i (k+1);

[0061] Among them, w is the inertia weight, c1 and c2 are acceleration factors, r1 and r2 are random numbers in [0,1], P i is the individual historical optimal position of the particle;

[0062] S44, based on the updated particle position X i (k+1) recalculate its fitness value F(X i (k+1)), and compare the fitness value with the fitness value of the reference solution F(X gbest ) for comparison:

[0063] If F(X i (k+1)) <F(P i ), then the individual historical optimal position of the updated particle is P i =X i (k+1);

[0064] If F(X i (k+1)) <F(X gbest )), then update the global optimal solution to X gbest =X i (k+1);

[0065] S45, loop through S41 to S44 until the chaotic particle swarm optimization algorithm meets the preset convergence conditions, including the maximum number of iterations K max Or the fitness value change is less than the threshold ∈, output the final global optimal solution X gbest1 .

[0066] Optionally, the S5 includes:

[0067] S51. Collect historical data and real-time input data of the charging pile operating environment, including user charging demand data D u , grid load change data D l and real-time electricity price fluctuation data D p ;

[0068] S52. Build a Bayesian inference model based on the historical data to perform prior probability modeling on the dynamic data of the charging pile operating environment, and set a prior distribution P(θ) of the model parameters, where θ is a set of model parameters including user charging demand parameters, grid load parameters, and electricity price fluctuation parameters;

[0069] S53. Using Bayes’ theorem combined with real-time input data D, update the posterior distribution P(θ|D) of the model parameters:

[0070]

[0071] Among them, P(θ|D) is the likelihood function, which means that the user charging demand data D is observed under the model parameter set θ. u , Grid load change data D l and real-time electricity price fluctuation data D p The probability of , P(D) is the normalization constant;

[0072] S54. User charging demand based on the updated posterior distribution P(θ|D) Grid load changes and real-time electricity price fluctuations Make a prediction:

[0073]

[0074]

[0075] Optionally, the S6 includes:

[0076] S61, introducing the charging pile operating environment prediction result generated by the Bayesian reasoning model as a dynamic input parameter into the charging pile multi-objective optimization model;

[0077] S62, dynamically adjust the target weight w in the multi-objective optimization model of the charging pile based on the prediction results of the charging pile operating environment.i And the constraints:

[0078]

[0079] Among them, α i is the sensitivity coefficient of target i, is the importance measure of target i;

[0080] The constraints are dynamically updated as follows:

[0081]

[0082] Among them, L t is the grid load at time t, L max The maximum load that the grid can bear, γ is the load margin adjustment coefficient, which is used to predict the value according to the grid load change Dynamically adjust grid load limits;

[0083] S63, using the updated target weight w i Recalculate the particle fitness value F based on the constraints new (X i ):

[0084]

[0085] in, Represents particle X i The fitness value under the energy consumption minimization goal, Represents particle X i The fitness value under the grid load balancing objective, Represents particle X i Fitness value under the goal of maximizing user satisfaction:

[0086]

[0087] in, Score the satisfaction of the i-th user:

[0088]

[0089] Among them, E i is the actual charging amount obtained by the i-th user, Forecast value based on user charging demand The minimum charging demand of the i-th user after dynamic adjustment;

[0090] S64, according to the updated fitness value F new (X i ) Sort each particle in the particle swarm and select the new current optimal particle Pnew and the global optimal particle X new ;

[0091] S65. Input the updated particle swarm state into the chaotic particle swarm optimization algorithm, and continue iterative optimization in combination with the chaotic perturbation mechanism until the algorithm meets the convergence condition.

[0092] An electronic device includes a charge and discharge control system, including the following modules:

[0093] The data acquisition module is used to collect historical data and real-time input data of the charging pile operating environment, including user charging demand data, grid load change data, and real-time electricity price fluctuation data;

[0094] The Bayesian reasoning module communicates with the data acquisition module to construct a priori probability distribution based on historical data, update the posterior distribution using real-time input data, and predict user charging demand, grid load changes, and real-time electricity price fluctuations based on the posterior distribution;

[0095] The multi-objective optimization module communicates with the Bayesian reasoning module and is used to dynamically adjust the objective weights and constraints based on the prediction results of the Bayesian reasoning module and calculate the fitness values of the particles to guide the optimization direction;

[0096] The chaotic particle swarm optimization module communicates with the multi-objective optimization module to initialize the particle swarm state, including the number of particles, initial position, and velocity. It iteratively updates the particle position and velocity, combines the chaotic perturbation mechanism to improve the global search capability, and optimizes the solution of the multi-objective charge-discharge model.

[0097] The charge and discharge control module communicates with the chaotic particle swarm optimization module, and is used to execute the charge and discharge strategy according to the global optimal solution output by the chaotic particle swarm optimization module, and provide real-time feedback on the execution effect;

[0098] The feedback adjustment module communicates with the charge and discharge control module to collect the actual effect of the charge and discharge control, dynamically update the input data of the Bayesian reasoning module and the target weight of the multi-objective optimization module, and form a closed-loop optimization control.

[0099] The beneficial effects of the present invention are:

[0100] (1) The present invention significantly improves the ability to solve multi-objective optimization problems by combining chaotic particle swarm optimization with Bayesian reasoning. The global search ability and local search efficiency of the algorithm are enhanced by introducing nonlinear chaotic mapping through chaotic particle swarm optimization, which solves the defect that traditional particle swarm optimization algorithm is prone to falling into local optimality in complex multi-objective scenarios. At the same time, combined with Bayesian reasoning, user charging demand, grid load changes and real-time electricity price fluctuations are dynamically predicted, and target weights and constraints are dynamically adjusted based on the prediction results to achieve balance and dynamic adaptation among multiple objectives. This enables the present invention to simultaneously meet the multiple requirements of minimizing energy consumption, balancing grid loads and maximizing user satisfaction in a changing operating environment, thereby improving the globality and adaptability of the optimization results.

[0101] (2) The present invention uses a Bayesian inference model to construct a dynamic prediction mechanism for the operating environment of a charging pile. It updates the posterior probability distribution based on historical data and real-time input data to generate prediction results of user charging demand, grid load changes, and electricity price fluctuations, and dynamically adjusts the optimization model based on these results. Compared with the traditional method of relying on static models for charging and discharging optimization, the dynamic prediction capability of the present invention significantly improves the adaptability of the charging pile to an uncertain environment, enabling the system to respond to changes in the external environment in real time, thereby maintaining efficient optimization control in complex scenarios. It is suitable for scenarios with large grid load fluctuations or frequent changes in user demand, and significantly improves the robustness and real-time performance of the system.

[0102] (3) The present invention improves the iterative update process of the traditional particle swarm optimization algorithm through the chaotic perturbation mechanism. The introduction of chaotic variables in the update of particle velocity and position makes the search trajectory of particles more diversified, significantly reducing the risk of the optimization process falling into the local optimum. In addition, the optimization direction is dynamically adjusted in combination with the prediction results of the Bayesian reasoning model, further improving the convergence speed and efficiency of the optimization process. Experiments show that the convergence time of the present invention in multi-objective high-dimensional optimization problems is shortened by about 30% compared with traditional methods, and the fitness value of the optimization result is significantly improved. It can quickly solve the global optimal solution in complex scenarios and meet the efficient operation requirements of electric vehicle charging piles in large-scale deployment scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0103] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0104] Figure 1 This is a flow chart of a charge and discharge control system and electronic equipment proposed by the present invention. DETAILED DESCRIPTION

[0105] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0106] refer to Figure 1 , a charge and discharge control system, comprising the following steps:

[0107] S1. Construct a multi-objective optimization model for charging piles, parameterize the multi-objective optimization problem into a multi-objective optimization model for charging piles, and define the objective function and constraints.

[0108] S2. Initialize the chaotic particle swarm optimization algorithm, set the number of particles, initial particle positions, particle velocity range, and chaos mapping function, generate an initial particle distribution with global search capability based on the chaos mapping, and use the initial particle distribution to solve the multi-objective optimization model of the charging pile;

[0109] S3. Calculate the fitness value of each particle in the particle swarm based on the objective function in the multi-objective optimization model of the charging pile, sort the particles according to the fitness value, select the current optimal particle and the global optimal particle, and store them as the initial candidate solutions of the global optimal solution;

[0110] S4. Using the initial candidate solution as the reference solution, combined with the chaotic perturbation mechanism, the current position and velocity of the particle are updated through a nonlinear chaotic mapping function. The updated particle state is used for the next fitness value calculation. The fitness of the reference solution is then compared and screened to enhance the particle swarm's ability to escape from the local optimal solution. This process is repeated until the chaotic particle swarm optimization algorithm reaches convergence.

[0111] S5. Use the Bayesian inference model to model and predict the dynamic data in the charging pile operating environment in real time, and generate prediction results of the charging pile operating environment;

[0112] S6. Introduce the charging pile operating environment prediction results generated by the Bayesian inference model into the chaotic particle swarm optimization algorithm, dynamically adjust the objective weights and constraints in the charging pile multi-objective optimization model, use the updated optimization model to recalculate the particle fitness values, and update the current optimal particle and the global optimal particle;

[0113] S7. Generate a multi-objective charging and discharging optimization control strategy for the charging pile based on the optimized global optimal particles, apply the optimization control strategy to the charging pile charging and discharging control execution module, and execute the multi-objective charging and discharging control of the charging pile.

[0114] In this embodiment, S1 includes:

[0115] S11. Construct a multi-objective optimization model for charging piles and parameterize the optimization objectives into objective functions, including the energy consumption minimization objective function minF1, the grid load balancing objective function minF2, and the user satisfaction maximization objective function minF3;

[0116] S12. Define the constraints of the multi-objective optimization model for charging piles, including:

[0117] Charging power constraints:

[0118] P min ≤P t ≤P max ;

[0119] Among them, P t is the charging power at time t, P min and P max are the minimum and maximum charging power allowed by the charging pile respectively;

[0120] Grid load limit:

[0121] L t ≤L max ;

[0122] Among them, L t is the grid load at time t, L max is the maximum load that the power grid can bear;

[0123] User charging requirements meet the following constraints:

[0124]

[0125] Among them, E i is the charging amount obtained by the i-th user during his charging time period, and are the charging start time and end time of the i-th user, The minimum charging amount required for the i-th user;

[0126] S13. Convert the multi-objective optimization problem into a multi-objective optimization model for charging piles, comprehensively consider the three objectives of minimizing energy consumption, balancing the grid load, and maximizing user satisfaction, and convert the multi-objective optimization problem into a single-objective optimization problem using the weighted summation method:

[0127] minF=w1minF1+w2minF2-w3minF3;

[0128] Among them, w1, w2, and w3 are target weight coefficients;

[0129] S14. Define the decision variables of the multi-objective optimization model of the charging pile as the charging power P in each time period. t.

[0130] In this embodiment, S2 includes:

[0131] S21, initialize the chaotic particle swarm optimization algorithm and set the number of particles N p , the initial position of the particle X i (0) and initial velocity V i (0), where i represents the i-th particle, X i The initial value of (0) is randomly distributed in the feasible solution space, V i The initial value of (0) is within the preset speed range [V min ,V max ] Randomly generated within;

[0132] S22. Define the chaotic mapping function C(k) using the Logistic mapping form:

[0133] C(k+1)=μC(k)(1-C(k));

[0134] Where k is the number of iterations, C(k) represents the chaotic variable, with an initial value of C(0)∈(0,1)), and μ is the control parameter of the chaotic mapping, with a value range of (0,4];

[0135] S23. Generate a chaotic sequence according to the chaotic mapping function, and use the chaotic sequence to adjust the initial position and initial velocity of the particle. The particle position is updated as follows:

[0136] X i (0) = X min +(X max -X min )·C(k);

[0137] The particle velocity is updated as:

[0138] V i (0) = V min +(V max -V min )·C(k);

[0139] Among them, X min and X max are the upper and lower bounds of the position, V min and V max are the upper and lower bounds of the velocity respectively;

[0140] S24, using the initialized particle position X i (0) and initial velocity V i (0) The initial particle distribution is introduced into the objective function of the multi-objective optimization model of the charging pile, and the fitness value F(X i(0)) and sort the particles by fitness, and select the initial global optimal particle G(0) and the initial individual optimal particle P i (0);

[0141] S25. Taking the initialized chaotic particle swarm state as input, an initial particle swarm with global search capability is formed.

[0142] In this embodiment, S3 includes:

[0143] S31. For each particle in the particle swarm, calculate the fitness value F(X i ):

[0144] F(X i )=w4F1(X i )+w5F2(X i )-w6F3(X i );

[0145] Among them, X i is the current position of the i-th particle, F1(X i )、F2(X i )、F3(X i ) are the energy consumption minimization objective function, the grid load balancing objective function and the user satisfaction maximization objective function at position X. i The value at , w4, w5, w6 are the target weight coefficients;

[0146] S32, based on the fitness value F(X i ) Sort the particles, the smaller the fitness value, the better the performance of the particle;

[0147] S33, record the individual historical optimal position P of each particle i And the corresponding optimal fitness value F(P i ), where P i is the position with the minimum fitness value found by the i-th particle so far;

[0148] S34. Select the particle with the smallest fitness value in the current particle swarm as the current global optimal particle G, and record the initial candidate solution X of the global optimal solution. gbest And the fitness value F gbest .

[0149] In this embodiment, S4 includes:

[0150] S41, the initial candidate solution X of the global optimal solution gbest As a reference solution, the current position X of the particle is calculated by combining the chaotic perturbation mechanism. i (k) and speed Vi (k) is updated, where k represents the current iteration number;

[0151] S42, generating a chaotic sequence C1(k) using a nonlinear chaotic mapping function;

[0152] S43. Update the particle velocity V based on the chaotic sequence C1(k) i (k+1) and position X i (k+1):

[0153] V i (k+1)=w·V i (k)+c1·r1·(P i -X i (k))+c2·r2·(X gbest -X i (k))+C1(k);

[0154] X i (k+1)=X i (k)+V i (k+1);

[0155] Among them, w is the inertia weight, c1 and c2 are acceleration factors, r1 and r2 are random numbers in [0,1], P i is the individual historical optimal position of the particle;

[0156] S44, based on the updated particle position X i (k+1) recalculate its fitness value F(X i (k+1)), and compare the fitness value with the fitness value of the reference solution F(X gbest ) for comparison:

[0157] If F(X i (k+1)) <F(P i ), then the individual historical optimal position of the updated particle is P i =X i (k+1);

[0158] If F(X i (k+1)) <F(X gbest )), then update the global optimal solution to X gbest =X i (k+1);

[0159] S45, loop through S41 to S44 until the chaotic particle swarm optimization algorithm meets the preset convergence conditions, including the maximum number of iterations K max Or the fitness value change is less than the threshold ∈, output the final global optimal solution X gbest1 .

[0160] In this embodiment, S5 includes:

[0161] S51. Collect historical data and real-time input data of the charging pile operating environment, including user charging demand data D u , Grid load change data D l and real-time electricity price fluctuation data D p ;

[0162] S52. Build a Bayesian inference model based on historical data to perform prior probability modeling on the dynamic data of the charging pile operating environment, and set the prior distribution P(θ) of the model parameters, where θ is a set of model parameters, including user charging demand parameters, grid load parameters, and electricity price fluctuation parameters;

[0163] S53. Using Bayes’ theorem combined with real-time input data D, update the posterior distribution P(θ|D) of the model parameters:

[0164]

[0165] Among them, P(θ|D) is the likelihood function, which means that the user charging demand data D is observed under the model parameter set θ. u , Grid load change data D l and real-time electricity price fluctuation data D p The probability of , P(D) is the normalization constant;

[0166] S54. User charging demand based on the updated posterior distribution P(θ|D) Grid load changes and real-time electricity price fluctuations Make a prediction:

[0167]

[0168] In this embodiment, S6 includes:

[0169] S61, introducing the charging pile operating environment prediction result generated by the Bayesian reasoning model as a dynamic input parameter into the charging pile multi-objective optimization model;

[0170] S62, dynamically adjust the target weight w in the multi-objective optimization model of the charging pile based on the prediction results of the charging pile operating environment. i And the constraints:

[0171]

[0172] Among them, α i is the sensitivity coefficient of target i, is the importance measure of target i;

[0173] The constraints are dynamically updated as follows:

[0174]

[0175] Among them, L t is the grid load at time t, L max The maximum load that the grid can bear, γ is the load margin adjustment coefficient, which is used to predict the value according to the grid load change Dynamically adjust grid load limits;

[0176] S63, using the updated target weight w i Recalculate the particle fitness value F based on the constraints new (X i ):

[0177]

[0178] in, Represents particle X i The fitness value under the energy consumption minimization goal, Represents particle X i The fitness value under the grid load balancing objective, Represents particle X i Fitness value under the goal of maximizing user satisfaction:

[0179]

[0180] in, Score the satisfaction of the i-th user:

[0181]

[0182] Among them, E i is the actual charging amount obtained by the i-th user, Forecast value based on user charging demand The minimum charging demand of the i-th user after dynamic adjustment;

[0183] S64, according to the updated fitness value F new (X i ) Sort each particle in the particle swarm and select the new current optimal particle P new and the global optimal particle X new ;

[0184] S65. Input the updated particle swarm state into the chaotic particle swarm optimization algorithm, and continue iterative optimization in combination with the chaotic perturbation mechanism until the algorithm meets the convergence condition.

[0185] An electronic device includes a charge and discharge control system, including the following modules:

[0186] The data acquisition module is used to collect historical data and real-time input data of the charging pile operating environment, including user charging demand data, grid load change data, and real-time electricity price fluctuation data;

[0187] The Bayesian reasoning module communicates with the data acquisition module to construct a priori probability distribution based on historical data, update the posterior distribution using real-time input data, and predict user charging demand, grid load changes, and real-time electricity price fluctuations based on the posterior distribution;

[0188] The multi-objective optimization module communicates with the Bayesian reasoning module and is used to dynamically adjust the objective weights and constraints based on the prediction results of the Bayesian reasoning module and calculate the fitness values of the particles to guide the optimization direction;

[0189] The chaotic particle swarm optimization module communicates with the multi-objective optimization module to initialize the particle swarm state, including the number of particles, initial position, and velocity. It iteratively updates the particle position and velocity, combines the chaotic perturbation mechanism to improve the global search capability, and optimizes the solution of the multi-objective charge-discharge model.

[0190] The charge and discharge control module communicates with the chaotic particle swarm optimization module, and is used to execute the charge and discharge strategy according to the global optimal solution output by the chaotic particle swarm optimization module, and provide real-time feedback on the execution effect;

[0191] The feedback adjustment module communicates with the charge and discharge control module to collect the actual effect of the charge and discharge control, dynamically update the input data of the Bayesian reasoning module and the target weight of the multi-objective optimization module, and form a closed-loop optimization control.

[0192] The present invention significantly improves the ability to solve multi-objective optimization problems by combining chaotic particle swarm optimization with Bayesian reasoning. The global search ability and local search efficiency of the algorithm are enhanced by introducing nonlinear chaotic mapping through chaotic particle swarm optimization, which solves the defect that traditional particle swarm optimization algorithms are prone to falling into local optimality in complex multi-objective scenarios. At the same time, combined with Bayesian reasoning, user charging needs, grid load changes and real-time electricity price fluctuations are dynamically predicted, and target weights and constraints are dynamically adjusted based on the prediction results to achieve balance and dynamic adaptation among multiple objectives. This enables the present invention to simultaneously meet the multiple needs of minimizing energy consumption, balancing grid loads and maximizing user satisfaction in a changing operating environment, thereby improving the globality and adaptability of the optimization results.

[0193] The present invention uses a Bayesian reasoning model to construct a dynamic prediction mechanism for the operating environment of a charging pile, updates the posterior probability distribution based on historical data and real-time input data to generate prediction results of user charging demand, grid load changes and electricity price fluctuations, and dynamically adjusts the optimization model based on these results. Compared with the traditional method of charging and discharging optimization that relies on static models, the dynamic prediction capability of the present invention significantly improves the adaptability of the charging pile to an uncertain environment, enabling the system to respond to changes in the external environment in real time, thereby maintaining efficient optimization control in complex scenarios. It is suitable for scenarios with large grid load fluctuations or frequent changes in user demand, and significantly improves the robustness and real-time performance of the system.

[0194] The present invention improves the iterative update process of the traditional particle swarm optimization algorithm through a chaotic perturbation mechanism. The introduction of chaotic variables in the update of particle velocity and position makes the particle search trajectory more diversified, significantly reducing the risk of the optimization process falling into the local optimum. In addition, the optimization direction is dynamically adjusted in combination with the prediction results of the Bayesian reasoning model, further improving the convergence speed and efficiency of the optimization process. Experiments show that the convergence time of the present invention in multi-objective high-dimensional optimization problems is shortened by about 30% compared with traditional methods, and the fitness value of the optimization result is significantly improved. It can quickly solve the global optimal solution in complex scenarios and meet the efficient operation requirements of electric vehicle charging piles in large-scale deployment scenarios.

[0195] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A charge and discharge control system, characterized in that: The steps include: S1. Construct a multi-objective optimization model for charging piles, parameterize the multi-objective optimization problem into a multi-objective optimization model for charging piles, and define the objective function and constraints. S2. Initialize the chaotic particle swarm optimization algorithm, set the number of particles, initial particle positions, particle velocity range, and chaos mapping function, generate an initial particle distribution with global search capability based on the chaos mapping, and use the initial particle distribution to solve the multi-objective optimization model of the charging pile; S3. Calculate the fitness value of each particle in the particle swarm based on the objective function in the multi-objective optimization model of the charging pile, sort the particles according to the fitness value, select the current optimal particle and the global optimal particle, and store them as the initial candidate solutions of the global optimal solution; S4. Using the initial candidate solution as the reference solution, combined with the chaotic perturbation mechanism, the current position and velocity of the particle are updated through a nonlinear chaotic mapping function. The updated particle state is used for the next fitness value calculation. The fitness of the reference solution is then compared and screened to enhance the particle swarm's ability to escape from the local optimal solution. This process is repeated until the chaotic particle swarm optimization algorithm reaches convergence. S5. Use the Bayesian inference model to model and predict the dynamic data in the charging pile operating environment in real time, and generate prediction results of the charging pile operating environment; S6. Introduce the charging pile operating environment prediction results generated by the Bayesian inference model into the chaotic particle swarm optimization algorithm, dynamically adjust the objective weights and constraints in the charging pile multi-objective optimization model, use the updated optimization model to recalculate the particle fitness values, and update the current optimal particle and the global optimal particle; S7. Generate a multi-objective charging and discharging optimization control strategy for the charging pile based on the optimized global optimal particle, apply the optimization control strategy to the charging pile charging and discharging control execution module, and execute the multi-objective charging and discharging control of the charging pile; Said S1 comprises: S11. Construct a multi-objective optimization model for charging piles and parameterize the optimization objectives into objective functions, including the energy consumption minimization objective function minF1, the grid load balancing objective function minF2, and the user satisfaction maximization objective function minF3; S12. Define the constraints of the multi-objective optimization model for charging piles, including: Charging power constraints: P min ≤P t ≤P max ; Among them, P t is the charging power at time t, P min and P max are the minimum and maximum charging power allowed by the charging pile respectively; Grid load limit: L t ≤L max ; Among them, L t is the grid load at time t, L max is the maximum load that the power grid can bear; User charging requirements meet the following constraints: Among them, E i is the charging amount obtained by the i-th user during his charging time period, and are the charging start time and end time of the i-th user, The minimum charging amount required for the i-th user; S13. Convert the multi-objective optimization problem into a multi-objective optimization model for charging piles, comprehensively consider the three objectives of minimizing energy consumption, balancing the grid load, and maximizing user satisfaction, and convert the multi-objective optimization problem into a single-objective optimization problem using the weighted summation method: minF=w1minF1+w2minF2-w3minF3; Among them, w1, w2, and w3 are target weight coefficients; S14. Define the decision variables of the multi-objective optimization model of the charging pile as the charging power P in each time period. t .

2. A charge and discharge control system according to claim 1, characterized in that: The S2 includes: S21, initialize the chaotic particle swarm optimization algorithm and set the number of particles N p , the initial position of the particle X i (0) and initial velocity V i (0), where i represents the i-th particle, X i The initial value of (0) is randomly distributed in the feasible solution space, V i The initial value of (0) is within the preset speed range [V min ,V max ] Randomly generated within; S22. Define the chaotic mapping function C(k) using the Logistic mapping form: C(k+1)=μC(k)(1-C(k)); Where k is the number of iterations, C(k) represents the chaotic variable, with an initial value of C(0)∈(0,1)), and μ is the control parameter of the chaotic mapping, with a value range of (0,4]; S23. Generate a chaotic sequence according to the chaotic mapping function, and use the chaotic sequence to adjust the initial position and initial velocity of the particle. The particle position is updated as follows: X i (0)=X min +(X max -X min )·C(k); The particle velocity is updated as: V i (0)=V min +(V max -V min )·C(k); Among them, X min and X max are the upper and lower bounds of the position, V min and V max are the upper and lower bounds of the velocity respectively; S24, using the initialized particle position X i (0) and initial velocity V i (0) The initial particle distribution is introduced into the objective function of the multi-objective optimization model of the charging pile, and the fitness value F(X i (0)) and sort the particles by fitness, and select the initial global optimal particle G(0) and the initial individual optimal particle P i (0); S25. Taking the initialized chaotic particle swarm state as input, an initial particle swarm with global search capability is formed.

3. A charge and discharge control system according to claim 1, characterized in that: The S3 includes: S31. For each particle in the particle swarm, calculate the fitness value F(X i ): F(X i )=w4F1(X i )+w5F2(X i )-w6F3(X i ); Among them, X i is the current position of the i-th particle, F1(X i )、F2(X i )、F3(X i ) are the energy consumption minimization objective function, the grid load balancing objective function and the user satisfaction maximization objective function at position X. i The value at , w4, w5, w6 are the target weight coefficients; S32, based on the fitness value F(X i ) Sort the particles, the smaller the fitness value, the better the performance of the particle; S33, record the individual historical optimal position P of each particle i And the corresponding optimal fitness value F(P i ), where P i is the position with the minimum fitness value found by the i-th particle so far; S34. Select the particle with the smallest fitness value in the current particle swarm as the current global optimal particle G, and record the initial candidate solution X of the global optimal solution. gbest And the fitness value F gbest .

4. A charge and discharge control system according to claim 1, characterized in that: The S4 includes: S41, the initial candidate solution X of the global optimal solution gbest As a reference solution, the current position X of the particle is calculated by combining the chaotic perturbation mechanism. i (k) and speed V i (k) is updated, where k represents the current iteration number; S42, generating a chaotic sequence C1(k) using a nonlinear chaotic mapping function; S43. Update the particle velocity V based on the chaotic sequence C1(k) i (k+1) and position X i (k+1): V i (k+1)=w·V i (k)+c1·r1·(P i -X i (k))+c2·r2·(X gbest -X i (k))+C1(k); X i (k+1)=X i (k)+V i (k+1); Among them, w is the inertia weight, c1 and c2 are acceleration factors, r1 and r2 are random numbers in [0,1], P i is the individual historical optimal position of the particle; S44, based on the updated particle position X i (k+1) recalculate its fitness value F(X i (k+1)), and compare the fitness value with the fitness value of the reference solution F(X gbest ) for comparison: If F(X i (k+1)) <F(P i ), then the individual historical optimal position of the updated particle is P i =X i (k+1); If F(X i (k+1))<F(X gbest )), then update the global optimal solution to X gbest =X i (k+1); S45, loop through S41 to S44 until the chaotic particle swarm optimization algorithm meets the preset convergence conditions, including the maximum number of iterations K max Or the fitness value change is less than the threshold ∈, output the final global optimal solution X gbest1 .

5. A charge and discharge control system according to claim 1, characterized in that: The S5 includes: S51. Collect historical data and real-time input data of the charging pile operating environment, including user charging demand data D u , grid load change data D l and real-time electricity price fluctuation data D p ; S52. Build a Bayesian inference model based on the historical data to perform prior probability modeling on the dynamic data of the charging pile operating environment, and set a prior distribution P(θ) of the model parameters, where θ is a set of model parameters including user charging demand parameters, grid load parameters, and electricity price fluctuation parameters; S53. Using Bayes’ theorem combined with real-time input data D, update the posterior distribution P(θ|D) of the model parameters: Among them, P(θ|D) is the likelihood function, which means that the user charging demand data D is observed under the model parameter set θ. u , grid load change data D l and real-time electricity price fluctuation data D p The probability of , P(D) is the normalization constant; S54. User charging demand based on the updated posterior distribution P(θ|D) Grid load changes and real-time electricity price fluctuations Make a prediction:

6. A charge and discharge control system according to claim 1, characterized in that: The S6 includes: S61, introducing the charging pile operating environment prediction result generated by the Bayesian reasoning model as a dynamic input parameter into the charging pile multi-objective optimization model; S62, dynamically adjust the target weight w in the multi-objective optimization model of the charging pile based on the prediction results of the charging pile operating environment. i And the constraints: Among them, α i is the sensitivity coefficient of target i, is the importance measure of target i; The constraints are dynamically updated as follows: Among them, L t is the grid load at time t, L max The maximum load that the grid can bear, γ is the load margin adjustment coefficient, which is used to predict the value according to the grid load change Dynamically adjust grid load limits; S63, using the updated target weight w i Recalculate the particle fitness value F based on the constraints new (X i ): in, Represents particle X i The fitness value under the energy consumption minimization goal, Represents particle X i The fitness value under the grid load balancing objective, Represents particle X i Fitness value under the goal of maximizing user satisfaction: in, Score the satisfaction of the i-th user: Among them, E i is the actual charging amount obtained by the i-th user, Forecast value based on user charging demand The minimum charging demand of the i-th user after dynamic adjustment; S64, according to the updated fitness value F new (X i ) Sort each particle in the particle swarm and select the new current optimal particle P new and the global optimal particle X new ; S65. Input the updated particle swarm state into the chaotic particle swarm optimization algorithm, and continue iterative optimization in combination with the chaotic perturbation mechanism until the algorithm meets the convergence condition.

7. An electronic device, characterized in that: A charge and discharge control system according to any one of claims 1 to 6, comprising the following modules: The data acquisition module is used to collect historical data and real-time input data of the charging pile operating environment, including user charging demand data, grid load change data, and real-time electricity price fluctuation data; The Bayesian reasoning module communicates with the data acquisition module to construct a priori probability distribution based on historical data, update the posterior distribution using real-time input data, and predict user charging demand, grid load changes, and real-time electricity price fluctuations based on the posterior distribution; The multi-objective optimization module communicates with the Bayesian reasoning module and is used to dynamically adjust the objective weights and constraints based on the prediction results of the Bayesian reasoning module and calculate the fitness values of the particles to guide the optimization direction; The chaotic particle swarm optimization module communicates with the multi-objective optimization module to initialize the particle swarm state, including the number of particles, initial position, and velocity. It iteratively updates the particle position and velocity, combines the chaotic perturbation mechanism to improve the global search capability, and optimizes the solution of the multi-objective charge-discharge model. The charge and discharge control module communicates with the chaotic particle swarm optimization module, and is used to execute the charge and discharge strategy according to the global optimal solution output by the chaotic particle swarm optimization module, and provide real-time feedback on the execution effect; The feedback adjustment module communicates with the charge and discharge control module to collect the actual effect of the charge and discharge control, dynamically update the input data of the Bayesian reasoning module and the target weight of the multi-objective optimization module, and form a closed-loop optimization control.

Images