A control method for photovoltaic vehicle charging station

Abstract

The present invention discloses a photovoltaic vehicle charging station control method, comprising step S1: combining a convolutional neural network in a machine learning algorithm with a gated cyclic unit to accurately predict the output power of the photovoltaic charging station; step S2: establishing an electric vehicle charging demand model; step S3: coordinating and controlling an energy storage system; and step S4: performing grid interaction and coordinated control, taking the maximum energy utilization efficiency as a global optimization goal. The photovoltaic vehicle charging station control method disclosed by the present invention solves the problems of low energy utilization, low charging efficiency, and poor grid stability in the original charging station.

Description

Technical Field

[0001] The invention belongs to the technical field of photovoltaic vehicle charging, and in particular relates to a photovoltaic vehicle charging station control method. Background Art

[0002] As new energy electric vehicles become more and more popular, more and more new energy vehicle charging stations are being built. Among them, photovoltaic vehicle charging stations, as a green and sustainable charging solution, have received more and more attention. However, the intermittent and volatile nature of photovoltaic energy generation and the uncertainty of electric vehicle charging demand have brought great challenges to the stable operation and efficient control of photovoltaic vehicle charging stations. Traditional control strategies often fail to give full play to the advantages of photovoltaic energy and cannot effectively meet the charging needs of electric vehicles.

[0003] Therefore, further improvements are made to the above problems. Summary of the invention

[0004] The main purpose of the present invention is to provide a photovoltaic vehicle charging station control method to solve the problems of low energy utilization, low charging efficiency, poor grid stability, etc. in the original charging station.

[0005] To achieve the above objectives, the present invention provides a photovoltaic vehicle charging station control method, comprising the following steps:

[0006] Step S1: combining the convolutional neural network (CNN) with the gated recurrent unit (GRU) in the machine learning algorithm to accurately predict the output power of the photovoltaic charging station;

[0007] Step S2: Establish an electric vehicle charging demand model (taking into account factors such as the battery capacity, remaining power, and charging time of the electric vehicle), predict the electric vehicle charging demand in different time periods, and use an intelligent optimization algorithm to schedule the charging process of the electric vehicle based on the predicted charging demand;

[0008] Step S3: Coordinated control of the energy storage system is performed, and the charging and discharging power of the energy storage system is adjusted in real time according to information including the output forecast value of photovoltaic power, the charging demand of electric vehicles and the state of the power grid (by reasonably controlling the charging and discharging power of the energy storage system, the fluctuation of photovoltaic power can be effectively smoothed, the energy utilization efficiency can be improved, and the stability of the power grid can be enhanced);

[0009] Step S4: Perform grid interaction and coordinated control, taking maximum energy utilization efficiency as the global optimization goal.

[0010] As a further preferred technical solution of the above technical solution, in step S1, the convolutional neural network is used to extract characteristic values ​​from meteorological data (light intensity, temperature, humidity, etc.), and the gated recurrent unit uses the extracted characteristic values ​​and historical power data to perform time prediction to obtain the output prediction value of photovoltaic power in the future period of time (when it is predicted that the photovoltaic power is sufficient, the charging demand of the electric vehicle is met first, and the excess electric energy is stored in the energy storage system; when it is predicted that the photovoltaic power is insufficient, the electric energy of the energy storage system and the power grid is reasonably dispatched to ensure the normal charging of the electric vehicle), which is specifically implemented as follows:

[0011] Step S1.1: Collect historical data of the photovoltaic charging station, including meteorological data, time information and corresponding photovoltaic power output data, remove outliers and reduce noise in the data, and normalize and pre-process the data;

[0012] Step S1.2: Construct a CNN model (convolutional neural network model), establish convolutional layers, activation functions, pooling layers and fully connected layers, take the preprocessed meteorological data as input, and the corresponding photovoltaic power output data as labels, train the CNN model, and extract (high-level) feature values;

[0013] Step S1.3: Construct a GRU time series architecture, stack multiple gated recurrent units together to form a deep GRU model (network), take the feature values ​​extracted by the CNN model and the historical photovoltaic power as input, divide the time window, train the GRU model, and then predict the future photovoltaic power based on the input feature values ​​and the historical photovoltaic power;

[0014] Step S1.4: Re-input meteorological data and historical photovoltaic power data into the trained CNN model and GRU model respectively, and convert the output x CNN With x GRU Splice to get x = [x CNN ;x GRU ], x CNN is the output element of the CNN model, x GRU is the output element of the GRU model;

[0015] Step S1.5: Introduce the attention mechanism, calculate the attention weight α, perform weighted calculation on x according to the attention weight vector α, and obtain the fused output intermediate variable x fused =α1x CNN +α2x GRU , and x fused Input into the fully connected layer, and finally get the output prediction value of the photovoltaic power of the photovoltaic charging station (in the prediction model combining CNN-GRU, the formula for the predicted photovoltaic power output value is finally obtained as follows in is the output prediction value of photovoltaic power, I is the light intensity, T is the temperature, RH is the humidity, is the historical photovoltaic data power).

[0016] As a further preferred technical solution of the above technical solution, step S2 is specifically implemented as the following steps:

[0017] Step S2.1: Establish optimization objectives, where:

[0018] Minimize charging time Where N is the number of electric vehicles, T i is the charging time of the i-th electric vehicle;

[0019] Balanced charging pile load Where M is the number of charging piles, L j is the load of the jth charging pile, is the average load of all charging piles;

[0020] Reduce grid load fluctuations Where T is the number of time steps, P grid (t) is the total load of the power grid at time t, is the average grid load;

[0021] Step S2.2: Establish constraints, where:

[0022] Electric vehicle battery capacity constraint: The charging capacity of each electric vehicle cannot exceed the battery capacity Where E i is the charging end time of the i-th electric vehicle, P i,t represents the charging power of the i-th electric vehicle at time t, Δt is the time step, C i is the battery capacity of the i-th electric car;

[0023] Charging pile power limit: The output power of each charging pile cannot exceed its rated power Among them I j is the set of electric vehicles connected to the jth charging station, P max,j is the power of the jth charging pile;

[0024] Grid power limit: The power drawn from the grid by the entire charging station cannot exceed the grid's allowable power. Where P pv (t) is the output power of the photovoltaic power station at time t, P grid,max is the maximum power allowed by the grid;

[0025] Charging continuity constraint: Once the car starts charging, it must continue charging until its demand is met or the maximum charging time is reached.i ≤t≤E i , then P i,t >0; otherwise P i,t =0, S i represents the charging start time of the i-th electric vehicle;

[0026] Time constraint: Charging start and end time must be within a reasonable time range, 0≤S i ≤T max , S i ≤E i ≤T max ;

[0027] Step S2.3: According to the established optimization objectives and constraints, the population is initialized by the genetic algorithm, and the optimization solution is obtained by the genetic algorithm to output the optimization parameters including the charging time and the power distribution.

[0028] As a further preferred technical solution of the above technical solution, step S3 is specifically implemented as the following steps:

[0029] Step S3.1: Establish optimization objectives, where:

[0030] Smoothing out PV power fluctuations (reducing the impact of PV power fluctuations on the grid and electric vehicle charging, with the goal of offsetting the output power of the energy storage system with the PV power fluctuations, making the total output power more stable) Where T is the number of time steps, P pv,t is the photovoltaic power at time t, P cha,t is the charging power of the energy storage system at time t, P discha,t is the discharge power of the energy storage system at time t, is the expected average total output power;

[0031] Improve energy efficiency (maximize the use of photovoltaic energy and energy storage system energy, reduce energy waste) Where P load,t is the electric vehicle charging power demand at time t;

[0032] Step S3.2: Establish constraints, where:

[0033] Energy storage capacity constraint: E min ≤E t ≤E max , where E t is the capacity of the energy storage system at time t, E min is the minimum capacity limit, E max is the maximum capacity limit, and the capacity time variation expression of the energy storage system is E t+1 =E t +(Pcha,t ×η cha -P discha,t / η discha )Δt, where η cha is the charging efficiency, η discha is the discharge efficiency, Δt is the time step;

[0034] Power limit constraint: The charging and discharging power of the energy storage system cannot exceed its maximum charging and discharging power limit in is the maximum discharge power, is the maximum charging power;

[0035] Interaction constraints between photovoltaic and grid: When photovoltaic power is in excess, the energy storage system can be charged, and the charging power does not exceed the difference between photovoltaic power generation and the charging demand of electric vehicles and the maximum power allowed by the grid. That is, if P pv,t -P load,t -P grid,max ≥0, then P cha,t ≤P pv,t -P load,t -P grid,max , where P grid,max is the maximum power allowed by the grid; when the photovoltaic power is insufficient, the energy storage system (can) discharge to meet the charging needs of electric vehicles, but the discharge power cannot exceed the remaining capacity of the energy storage system and the charging needs of electric vehicles, that is, if P pv,t -P load,t <0, then P discha,t ≤min(E t ,P load,t -P pv,t );

[0036] Continuity constraint: The charge and discharge state of the energy storage system cannot change suddenly between adjacent time steps to ensure the stability of the system. cha,t >0, and P cha,t-1 =0, then If P discha,t >0 and P discha,t-1 =0, then in and is the maximum allowable charge and discharge power variation;

[0037] Step S3.3: According to the established optimization objectives and constraints, the population is initialized by the genetic algorithm, and the optimization solution is obtained by the genetic algorithm to output the optimization parameters including the charging and discharging power values ​​and the energy storage state.

[0038] As a further preferred technical solution of the above technical solution, for step S4:

[0039] (The photovoltaic charging station should interact and coordinate well with the power grid to achieve efficient use of energy and stable operation of the power grid) Establish a communication interface between the photovoltaic charging station and the power grid, monitor the parameters of the power grid including voltage, frequency, and power in real time, and feed back the operating status information of the photovoltaic charging station to the power grid;

[0040] According to the demand and operation status of the power grid, a distributed control strategy is used to adjust the power output and charging strategy of the photovoltaic charging station. The optimization goal of the entire system is Where n is the number of subsystems, f i (x i ) is the objective function of each subsystem i, x i is the decision variable of subsystem i. Communication interfaces are implanted between systems to ensure information exchange between systems. Each subsystem exchanges information with neighboring subsystems to gradually optimize its own decision variables, so that the objective function of the entire system is minimized. The corresponding actual situation is to maximize energy efficiency as the global optimization goal. Where E out With E in are the total output and input energy of the system in the time interval [t1, t2] respectively.

[0041] The beneficial effects of the present invention are:

[0042] 1. Power prediction method

[0043] The photovoltaic power is predicted by combining convolutional neural network (CNN) with gated recurrent unit (GRU). CNN is used to extract meteorological data features, and GRU uses the extracted features and historical power data for time series prediction. This combination makes full use of the advantages of CNN in feature extraction and GRU in time series processing, improving the accuracy of prediction.

[0044] By introducing the attention mechanism to fuse the outputs of CNN and GRU, the model can automatically focus on the information that is more important to the prediction results, further improving the prediction effect.

[0045] 2. Scheduling optimization among systems

[0046] The electric vehicle charging subsystem and energy storage subsystem are modeled separately, and genetic algorithms are used to optimize the subsystems. The charging and discharging power of the energy storage system is adjusted in real time according to the predicted photovoltaic power, electric vehicle charging demand and grid status to improve the charging efficiency and system stability.

[0047] 3. Reasonable grid interaction and coordinated control

[0048] A communication interface between the photovoltaic vehicle charging station and the power grid was established, which enabled real-time monitoring and information feedback, and provided a basis for coordinated control.

[0049] 4. A distributed control strategy is adopted, with maximizing energy utilization efficiency as the global optimization goal. The decision variables are gradually optimized through information exchange between subsystems, thus achieving efficient operation of the entire system. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is the overall flow chart of the control strategy of the present invention.

[0051] Figure 2 It is a structural diagram of a prediction model combining a convolutional neural network and a gated recurrent unit of the present invention.

[0052] Figure 3 It is a system architecture diagram of the photovoltaic charging station of the present invention. DETAILED DESCRIPTION

[0053] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0054] In the preferred embodiments of the present invention, those skilled in the art should note that the photovoltaic charging station and the like involved in the present invention may be regarded as prior art.

[0055] Preferred embodiments.

[0056] like Figure 1-3 As shown, the present invention discloses a photovoltaic vehicle charging station control method, comprising the following steps:

[0057] Step S1: combining the convolutional neural network (CNN) with the gated recurrent unit (GRU) in the machine learning algorithm to accurately predict the output power of the photovoltaic charging station;

[0058] Step S2: Establish an electric vehicle charging demand model (taking into account factors such as the battery capacity, remaining power, and charging time of the electric vehicle), predict the electric vehicle charging demand in different time periods, and use an intelligent optimization algorithm to schedule the charging process of the electric vehicle based on the predicted charging demand;

[0059] Step S3: Coordinated control of the energy storage system is performed, and the charging and discharging power of the energy storage system is adjusted in real time according to information including the output forecast value of photovoltaic power, the charging demand of electric vehicles and the state of the power grid (by reasonably controlling the charging and discharging power of the energy storage system, the fluctuation of photovoltaic power can be effectively smoothed, the energy utilization efficiency can be improved, and the stability of the power grid can be enhanced);

[0060] Step S4: Perform grid interaction and coordinated control, taking maximum energy utilization efficiency as the global optimization goal.

[0061] Specifically, in step S1, the convolutional neural network is used to extract characteristic values ​​from meteorological data (light intensity, temperature, humidity, etc.), and the gated recurrent unit uses the extracted characteristic values ​​and historical power data to perform time prediction to obtain the output prediction value of photovoltaic power in the future (when it is predicted that the photovoltaic power is sufficient, the charging demand of electric vehicles is met first, and the excess electric energy is stored in the energy storage system; when it is predicted that the photovoltaic power is insufficient, the energy storage system and the power grid are reasonably dispatched to ensure the normal charging of electric vehicles). The specific implementation is the following steps:

[0062] Step S1.1: Collect historical data of the photovoltaic charging station, including meteorological data, time information and corresponding photovoltaic power output data, remove outliers and reduce noise in the data, and normalize and pre-process the data;

[0063] Step S1.2: Construct a CNN model (convolutional neural network model), establish convolutional layers, activation functions, pooling layers and fully connected layers, take the preprocessed meteorological data as input, and the corresponding photovoltaic power output data as labels, train the CNN model, and extract (high-level) feature values;

[0064] Step S1.3: Construct a GRU time series architecture, stack multiple gated recurrent units together to form a deep GRU model (network), take the feature values ​​extracted by the CNN model and the historical photovoltaic power as input, divide the time window, train the GRU model, and then predict the future photovoltaic power based on the input feature values ​​and the historical photovoltaic power;

[0065] Step S1.4: Re-input meteorological data and historical photovoltaic power data into the trained CNN model and GRU model respectively, and convert the output x CNN With x GRU Splice to get x = [x CNN ;x GRU ], x CNN is the output element of the CNN model, x GRU is the output element of the GRU model;

[0066] Step S1.5: Introduce the attention mechanism, calculate the attention weight α, perform weighted calculation on x according to the attention weight vector α, and obtain the fused output intermediate variable x fused =α1x CNN +α2x GRU , and x fused Input into the fully connected layer, and finally get the output prediction value of the photovoltaic power of the photovoltaic charging station (in the prediction model combining CNN-GRU, the formula for the predicted photovoltaic power output value is finally obtained as follows in is the output prediction value of photovoltaic power, I is the light intensity, T is the temperature, RH is the humidity, is the historical photovoltaic data power).

[0067] More specifically, step S2 is implemented as follows:

[0068] Step S2.1: Establish optimization objectives, where:

[0069] Minimize charging time Where N is the number of electric vehicles, T i is the charging time of the i-th electric vehicle;

[0070] Balanced charging pile load Where M is the number of charging piles, L j is the load of the jth charging pile, is the average load of all charging piles;

[0071] Reduce grid load fluctuations Where T is the number of time steps, P grid (t) is the total load of the power grid at time t, is the average grid load;

[0072] Step S2.2: Establish constraints, where:

[0073] Electric vehicle battery capacity constraint: The charging capacity of each electric vehicle cannot exceed the battery capacity Where E i is the charging end time of the i-th electric vehicle, P i,t represents the charging power of the i-th electric vehicle at time t, Δt is the time step, C i is the battery capacity of the i-th electric car;

[0074] Charging pile power limit: The output power of each charging pile cannot exceed its rated power Among them I j is the set of electric vehicles connected to the jth charging station, P max,jis the power of the jth charging pile;

[0075] Grid power limit: The power drawn from the grid by the entire charging station cannot exceed the grid's allowable power. Where P pv (t) is the output power of the photovoltaic power station at time t, P grid,max is the maximum power allowed by the grid;

[0076] Charging continuity constraint: Once the car starts charging, it must continue charging until its demand is met or the maximum charging time is reached. i ≤t≤E i , then P i,t >0; otherwise P i,t =0, S i represents the charging start time of the i-th electric vehicle;

[0077] Time constraint: Charging start and end time must be within a reasonable time range, 0≤S i ≤T max , S i ≤E i ≤T max ;

[0078] Step S2.3: According to the established optimization objectives and constraints, the population is initialized by the genetic algorithm, and the genetic algorithm is optimized to output the optimization parameters including charging time and power distribution, which is specifically implemented as follows:

[0079] The charging start time and charging power at each moment of each electric vehicle are encoded and processed, and the initial population is randomly generated; the fitness function value is calculated according to the optimization target, and the individuals in the population are sorted according to the fitness function value; the parent individual is selected from the current population for crossover operation; the mutation probability is set, and each gene in the offspring individual is judged whether to mutate according to the mutation probability; whether to stop iteration is judged according to the set termination condition; finally, the optimal individual is decoded to obtain the charging start time and charging power at each moment of each electric vehicle, and the optimization parameters of the charging time and power allocation of each electric vehicle are calculated.

[0080] Furthermore, step S3 is specifically implemented as the following steps:

[0081] Step S3.1: Establish optimization objectives, where:

[0082] Smoothing out PV power fluctuations (reducing the impact of PV power fluctuations on the grid and electric vehicle charging, with the goal of offsetting the output power of the energy storage system with the PV power fluctuations, making the total output power more stable) Where T is the number of time steps, P pv,tis the photovoltaic power at time t, P cha,t is the charging power of the energy storage system at time t, P discha,t is the discharge power of the energy storage system at time t, is the expected average total output power;

[0083] Improve energy efficiency (maximize the use of photovoltaic energy and energy storage system energy, reduce energy waste) Where P load,t is the electric vehicle charging power demand at time t;

[0084] Step S3.2: Establish constraints, where:

[0085] Energy storage capacity constraint: E min ≤E t ≤E max , where E t is the capacity of the energy storage system at time t, E min is the minimum capacity limit, E max is the maximum capacity limit, and the capacity time variation expression of the energy storage system is E t+1 =E t +(P cha,t ×η cha -P discha,t / η discha )Δt, where η cha is the charging efficiency, η discha is the discharge efficiency, Δt is the time step;

[0086] Power limit constraint: The charging and discharging power of the energy storage system cannot exceed its maximum charging and discharging power limit in is the maximum discharge power, is the maximum charging power;

[0087] Interaction constraints between photovoltaic and grid: When photovoltaic power is in excess, the energy storage system can be charged, and the charging power does not exceed the difference between photovoltaic power generation and the charging demand of electric vehicles and the maximum power allowed by the grid. That is, if P pv,t -P load,t -P grid,max ≥0, then P cha,t ≤P pv,t -P load,t -P grid,max , where P grid,max is the maximum power allowed by the grid; when the photovoltaic power is insufficient, the energy storage system (can) discharge to meet the charging needs of electric vehicles, but the discharge power cannot exceed the remaining capacity of the energy storage system and the charging needs of electric vehicles, that is, if P pv,t -P load,t <0, then Pdischa,t ≤min(E t ,P load,t -P pv,t );

[0088] Continuity constraint: The charge and discharge state of the energy storage system cannot change suddenly between adjacent time steps to ensure the stability of the system. cha,t >0, and P cha,t-1 =0, then If P discha,t >0 and P discha,t-1 =0, then in and is the maximum allowable charge and discharge power variation;

[0089] Step S3.3: According to the established optimization objectives and constraints, the population is initialized by the genetic algorithm, and the optimization solution is obtained by the genetic algorithm to output the optimization parameters including the charging and discharging power values ​​and the energy storage state, which is specifically implemented as follows:

[0090] The charging and discharging power is encoded, and the initial population is randomly generated. The chromosomes of the individuals are composed of the charging and discharging power codes at each moment. The fitness function is designed according to the optimization objectives and constraints, and the weighted combination of the target value of smoothing the photovoltaic power fluctuation and the target value of improving the energy utilization efficiency is calculated to obtain the fitness function value. The parent individuals are selected from the current population for crossover operation until a sufficient number of parent individuals are selected. The mutation probability is set, and each gene in the offspring individual is judged whether to perform a mutation operation on it according to the mutation probability. Whether to stop the iteration is determined according to the set termination conditions. The optimal individual is decoded to obtain the charging and discharging power at each moment, and the charging and discharging power value at each moment and the optimization parameters such as the energy storage state are calculated and output.

[0091] Furthermore, for step S4:

[0092] (The photovoltaic charging station should interact and coordinate well with the power grid to achieve efficient use of energy and stable operation of the power grid) Establish a communication interface between the photovoltaic charging station and the power grid, monitor the parameters of the power grid including voltage, frequency, and power in real time, and feed back the operating status information of the photovoltaic charging station to the power grid;

[0093] According to the demand and operation status of the power grid, a distributed control strategy is used to adjust the power output and charging strategy of the photovoltaic charging station. The optimization goal of the entire system is Where n is the number of subsystems, f i (x i ) is the objective function of each subsystem i, x iis the decision variable of subsystem i. Communication interfaces are implanted between systems to ensure information exchange between systems. Each subsystem exchanges information with neighboring subsystems to gradually optimize its own decision variables, so that the objective function of the entire system is minimized. The corresponding actual situation is to maximize energy efficiency as the global optimization goal. Where E out With E in are the total output and input energy of the system in the time interval [t1, t2] respectively.

[0094] It is worth mentioning that the technical features such as photovoltaic charging stations involved in the patent application of this invention should be regarded as prior art. The specific structure, working principle and possible control method and spatial layout method of these technical features can be selected according to the conventional selection in the field, and should not be regarded as the inventive point of the patent of this invention. The patent of this invention will not be further elaborated.

[0095] For those skilled in the art, it is still possible to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A photovoltaic vehicle charging station control method, characterized in that: The following steps are involved: Step S1: combining the convolutional neural network in the machine learning algorithm with the gated recurrent unit to accurately predict the output power of the photovoltaic charging station; Step S2: Establishing an electric vehicle charging demand model, predicting the electric vehicle charging demand in different time periods, and scheduling the charging process of the electric vehicle using an intelligent optimization algorithm based on the predicted charging demand; Step S3: Coordinated control of the energy storage system, and real-time adjustment of the charging and discharging power of the energy storage system based on information including the output forecast value of photovoltaic power, electric vehicle charging demand, and grid status; Step S4: Conduct grid interaction and coordinated control, taking maximum energy utilization efficiency as the global optimization goal.

2. A photovoltaic vehicle charging station control method according to claim 1, characterized in that: In step S1, the convolutional neural network is used to extract the characteristic values ​​in the meteorological data, and the gated recurrent unit uses the extracted characteristic values ​​and historical power data to perform time prediction to obtain the output prediction value of the photovoltaic power in the future period of time, which is specifically implemented as follows: Step S1.1: Collect historical data of the photovoltaic charging station, including meteorological data, time information and corresponding photovoltaic power output data, remove outliers and reduce noise in the data, and normalize and pre-process the data; Step S1.2: Construct a CNN model, establish convolutional layers, activation functions, pooling layers, and fully connected layers, use the preprocessed meteorological data as input, and the corresponding photovoltaic power output data as labels, train the CNN model, and extract feature values; Step S1.3: Construct a GRU time series architecture, stack multiple gated recurrent units together to form a deep GRU model, take the feature values ​​extracted by the CNN model and the historical photovoltaic power as input, divide the time window, train the GRU model, and then predict the future photovoltaic power based on the input feature values ​​and the historical photovoltaic power; Step S1.4: Re-input meteorological data and historical photovoltaic power data into the trained CNN model and GRU model respectively, and convert the output x CNN With x GRU Splice to get x = [x CNN ;x GRU ], x CNN is the output element of the CNN model, x GRU is the output element of the GRU model; Step S1.5: Introduce the attention mechanism, calculate the attention weight α, perform weighted calculation on x according to the attention weight vector α, and obtain the fused output intermediate variable x fused =α1x CNN +α2x GRU , and x fused It is input into the fully connected layer, and finally the output prediction value of the photovoltaic power of the photovoltaic charging station is obtained.

3. A photovoltaic vehicle charging station control method according to claim 2, characterized in that: Step S2 is specifically implemented as the following steps: Step S2.1: Establish optimization objectives, where: Minimize charging time Where N is the number of electric vehicles, T i is the charging time of the i-th electric vehicle; Balanced charging pile load Where M is the number of charging piles, L j is the load of the jth charging pile, is the average load of all charging piles; Reduce grid load fluctuations Where T is the number of time steps, P grid (t) is the total load of the power grid at time t, is the average grid load; Step S2.2: Establish constraints, where: Electric vehicle battery capacity constraint: The charging capacity of each electric vehicle cannot exceed the battery capacity Where E i is the charging end time of the i-th electric vehicle, P i,t represents the charging power of the i-th electric vehicle at time t, Δt is the time step, C i is the battery capacity of the i-th electric car; Charging pile power limit: The output power of each charging pile cannot exceed its rated power Among them I j is the set of electric vehicles connected to the jth charging station, P max,j is the power of the jth charging pile; Grid power limit: The power drawn from the grid by the entire charging station cannot exceed the grid's allowable power. Where P pv (t) is the output power of the photovoltaic power station at time t, P grid,max is the maximum power allowed by the grid; Charging continuity constraint: Once the car starts charging, it must continue charging until its demand is met or the maximum charging time is reached. i ≤t≤E i , then P i,t >0; otherwise P i,t =0, S i represents the charging start time of the i-th electric vehicle; Time constraint: Charging start and end time must be within a reasonable time range, 0≤S i ≤T max , S i ≤E i ≤T max ; Step S2.3: According to the established optimization objectives and constraints, the population is initialized by the genetic algorithm, and the optimization solution is obtained by the genetic algorithm to output the optimization parameters including the charging time and the power distribution.

4. A photovoltaic vehicle charging station control method according to claim 3, characterized in that: Step S3 is specifically implemented as the following steps: Step S3.1: Establish optimization objectives, where: Smoothing out photovoltaic power fluctuations Where R is the number of time steps, P pv,t is the photovoltaic power at time t, P cha,t is the charging power of the energy storage system at time t, P discha,t is the discharge power of the energy storage system at time t, is the expected average total output power; Improve energy efficiency Where P load,t is the electric vehicle charging power demand at time t; Step S3.2: Establish constraints, where: Energy storage capacity constraint: E min ≤E t ≤E max , where E t is the capacity of the energy storage system at time t, E min is the minimum capacity limit, E max is the maximum capacity limit, and the capacity time variation expression of the energy storage system is E t+1 =E t +(P cha,t ×η cha -P discha,t / η discha )Δt, where η cha is the charging efficiency, η discha is the discharge efficiency, Δt is the time step; Power limit constraint: The charging and discharging power of the energy storage system cannot exceed its maximum charging and discharging power limit in is the maximum discharge power, is the maximum charging power; Interaction constraints between photovoltaic and grid: When photovoltaic power is in excess, the energy storage system is charged, and the charging power does not exceed the difference between photovoltaic power generation and the charging demand of electric vehicles and the maximum power allowed by the grid. That is, if P pv,t -P load,t -P grid,max ≥0, then P cha,t ≤P pv,t -P load,t -P grid,max , where P grid,max is the maximum power allowed by the grid; when the photovoltaic power is insufficient, the energy storage system discharges to meet the charging needs of electric vehicles, but the discharge power cannot exceed the remaining capacity of the energy storage system and the charging needs of electric vehicles, that is, if P pv,t -P load,t <0, then P discha,t ≤min(E t ,P load,t -P pv,t ); Continuity constraint: The charge and discharge state of the energy storage system cannot change suddenly between adjacent time steps to ensure the stability of the system. cha,t >0, and P cha,t-1 =0, then If P discha,t >0 and P discha,t-1 =0, then in and is the maximum allowable charge and discharge power variation; Step S3.3: According to the established optimization objectives and constraints, the population is initialized by the genetic algorithm, and the genetic algorithm is optimized and solved to output the optimization parameters including the charging and discharging power values ​​and the energy storage state.

5. A photovoltaic vehicle charging station control method according to claim 4, characterized in that: For step S4: Establish a communication interface between the photovoltaic charging station and the power grid, monitor the parameters of the power grid including voltage, frequency and power in real time, and feed back the operating status information of the photovoltaic charging station to the power grid; According to the demand and operation status of the power grid, a distributed control strategy is used to adjust the power output and charging strategy of the photovoltaic charging station. The optimization goal of the entire system is Where n is the number of subsystems, f i (x i ) is the objective function of each subsystem i, x i is the decision variable of subsystem i. Communication interfaces are implanted between systems to ensure information exchange between systems. Each subsystem exchanges information with neighboring subsystems to gradually optimize its own decision variables, so that the objective function of the entire system is minimized. The corresponding actual situation is to maximize energy efficiency as the global optimization goal. Where E out With E in are the total output and input energy of the system in the time interval [t1, t2] respectively.

Images