Quick charging station energy-saving control system and method with peak-valley transfer function

Abstract

The invention discloses a quick charging station energy-saving control system and method with a peak-valley transfer function, and relates to the field of charging station optimization. An induction communication module induces the access of an electric vehicle, stores a corresponding charging record and updates the residual electric quantity of other accessed electric vehicles; the power grid monitoring module determines a global suspension time period based on the charging record, generates a predicted electricity price sequence by adopting a time sequence prediction network, and checks whether a predicted error decision of each time period regenerates the predicted electricity price sequence or not in the charging process; the charging planning module calls a predicted electricity price sequence, obtains a charging record of an accessed electric vehicle, sets constraints by considering battery health, electricity price change and power grid load safety, combines and constructs a cost optimization model with the goal of minimizing global electricity cost, introduces a Lagrangian multiplier and rewrites the same into an unconstrained optimization problem, and finally performs charging planning. And an improved mouse swarm optimization algorithm is adopted to solve, extract and restore an optimal charging power sequence group, so that self-adaptive charging planning considering cost, safety and battery health is realized.

Description

Technical Field

[0001] The present invention relates to the field of charging station optimization, and in particular to a fast charging station energy-saving control system and method with a peak-valley load shifting function. Background Art

[0002] In order to balance the load of the power grid and improve energy efficiency, how to take advantage of the difference in peak and valley electricity prices and reasonably arrange the charging time and power of electric two-wheelers and three-wheelers has become an urgent problem to be solved by charging stations.

[0003] The existing Chinese patent application with publication number CN117293873A discloses an online optimization system for charging and discharging electric vehicles for shifting peak loads and filling valley loads, including a data acquisition module and an online optimization module; the data acquisition module is used to collect data; the online optimization module is used to establish a system model and optimization objectives, and then based on a rolling optimization algorithm, the electric vehicle status data is input to obtain a sliding window of the vehicle charging time, and a distributed alternating direction multiplier algorithm is executed within the range of the sliding window. In the distributed alternating direction multiplier algorithm, the electric vehicle status data and charging station load data collected by the data acquisition module are input to obtain the optimal charging and discharging power of each electric vehicle.

[0004] Existing charging optimization technology only optimizes the charging station load to minimize the target, ignoring the price changes of grid electricity prices and the health factor of electric vehicle batteries, and is unable to cope with complex electric vehicle charging planning scenarios. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention proposes an energy-saving control system and method for a fast charging station with a peak-valley load shifting function, which comprehensively considers the health factor of the battery and the changing trend of the power grid electricity price, and dynamically plans the charging power with the lowest electricity cost as the goal.

[0006] The technical solution to achieve the purpose of the present invention is: The energy-saving control system of the fast charging station with peak-valley load shifting function includes an inductive communication module, a power grid monitoring module and a charging planning module; The inductive communication module senses the Access electric vehicles and obtain charging records And save it in the charging queue list to obtain the connected The remaining power of the electric vehicle is updated to update the charging queue list and broadcast the planning command, sensing the When an electric vehicle is disconnected, the corresponding charging record is deleted And broadcast vehicle clearing command, charging record Include battery information and charging demand information , , is the total number of ports; The grid monitoring module receives the planning command, queries the charging queue table to determine the global suspension period, and uses the time series prediction network to generate the predicted electricity price sequence , check the prediction error in each period before the global suspension period, decide whether to broadcast the electricity price fluctuation command and regenerate the predicted electricity price sequence ; The charging planning module receives the planning command or the electricity price fluctuation command and calls the predicted electricity price sequence , query the charging queue table to obtain the rated power, charging lower limit power, charging upper limit power, remaining power, health coefficient and expected end time of each electric vehicle, consider the health coefficient of the electric vehicle and the trend of electricity price changes to build a cost optimization model and rewrite it into an unconstrained optimization problem by introducing Lagrange multipliers, use the improved rat swarm optimization algorithm to solve the unconstrained optimization problem to generate the optimal solution, extract and restore it to the optimal charging power sequence group and execute it, receive the vehicle clearing command, retrieve the optimal charging power sequence group to decide whether to clear the corresponding vehicle optimal charging power sequence.

[0007] Further, the inductive communication module includes a port detection unit, a two-way communication unit and a queue storage unit; The port detection unit monitors based on hardware detection circuits The detection current of each port. When the detection current of the port changes from 0 to non-0, it indicates that the electric vehicle is connected to the port. When the detection current of the port changes from non-0 to 0, it indicates that the electric vehicle is disconnected from the port. is the total number of ports; The dual-channel communication unit is based on the CAN communication protocol and The battery management system of the connected electric vehicle establishes a wired link to obtain battery information , based on wireless communication technology and The mobile device used by the user of an electric vehicle establishes a wireless link to obtain charging demand information , combined to generate charging records , through the established Before the wired link is reacquired The remaining power of an electric vehicle, including battery information Including rated power , Charging lower limit power , Charging upper limit power , remaining power and health factor , charging demand information Including charging start time and expected end time ; Before the queue storage unit is acquired The remaining power of the electric vehicle is replaced with the remaining power recorded in the charging queue list to obtain the Electric vehicle charging records And store in the charging queue list, broadcast planning commands, when the When an electric vehicle is disconnected, the charging record is deleted And broadcast the vehicle clearance command.

[0008] Furthermore, the power grid monitoring module includes a power grid connection unit, an electricity price prediction unit and a time period verification unit; The grid connection unit obtains the electricity price in each time period and stores it in the historical time period electricity price table in sequence; The electricity price prediction unit receives the planning command or correction signal, queries the charging queue table, takes the time period of the last expected end time in the existing charging record as the global suspension period, and uses the time series prediction network combined with the historical time period electricity price table to generate a predicted electricity price sequence. , if no new planning command is received at the end of the global suspension period, the predicted electricity price sequence is cleared; The time period verification unit checks whether there is a predicted electricity price sequence in each time period. If it does not exist, no processing will be done. If it does exist, the period electricity price and period predicted electricity price corresponding to the current period will be called and the prediction error will be calculated. Based on the comparison result between the prediction error and the upper limit of the error, it will be decided whether to send a correction signal to the electricity price prediction unit and broadcast the electricity price fluctuation command.

[0009] Furthermore, a time series forecasting network is used to generate a forecasted electricity price sequence The specific steps include: Count the total number of time periods from the current time period to the global termination time period And mark the time periods in sequence as ; Extract the electricity price closest to the current period from the historical period electricity price table The electricity price for each historical period , construct the electricity price series for historical periods ; Normalized historical period electricity price series And through one-dimensional convolution and ReLU activation function, feature extraction is performed to generate a historical electricity price representation sequence ; Capturing historical electricity price representation sequences through temporal attention layer The single electricity price characterization and other The time dependence of the electricity price representation is adaptively assigned to the rest The different attention weights of the electricity price representations are weighted to re-represent a single electricity price representation. The power price representations are re-represented to obtain the weighted power price representation sequence ; Capturing weighted price representation sequences through gate-controlled recurrent layers The hidden state matrix of a single period in the process is updated to generate the hidden state matrix of the next period, and the hidden state matrix of the historical period with period number 0 is obtained by looping. The hidden state matrix of the historical period numbered 0 is flattened by a flattening layer, dimension reduction is performed through a fully connected layer, and anti-minimum-maximum normalization is used to restore it to the predicted electricity price sequence. .

[0010] Furthermore, the pre-training of the time series prediction network is also achieved by improving the rat swarm optimization algorithm, including the following specific steps: The time series prediction network The hyperparameters are organized into a hyperparameter vector and constructed dimensional vector space and regard the hyperparameter vector as the position in the vector space. The evaluation index is defined as the error between the electricity price in the predicted period and the electricity price in the actual period. is the total number of hyperparameters; In the 0th iteration, Random initialization in a vector space of A mouse, The positions of the mice are used as hyperparameters of the time series prediction network to calculate the evaluation index and determine the optimal position of round 0. , is the total number of mice; Set the maximum number of iterations , in In the round iteration, , based on Optimal wheel position Determine The leader mouse of the iteration, the rest The servant mouse from The position of the wheel is randomly explored. The random exploration direction of the servant mouse is determined by The servant mouse The position of the wheel points to Optimal wheel position , the distance of random exploration increases with the number of iterations decreases with the increase of , and introduces random numbers to increase randomness; Randomly select multiple groups of non-repetitive servant mouse pairs, cross-derive the two servant mice in each group of servant mouse pairs, derive the servant mouse at the new position and randomly replace one servant mouse in the servant mouse pair through the linear combination of the positions of the two servant mice or the random exchange of elements of different dimensions in the positions; The first Round Iteration The position of the mouse is used as the hyperparameter of the time series prediction network to calculate the evaluation index and determine the Optimal wheel position ; Start the next round of iteration and perform random exploration and cross-derivation until the maximum number of iterations is reached ,at this time The location where the mice gather is the optimal location , the best position of The hyperparameters are the hyperparameters when the time series prediction network pre-training is completed.

[0011] Furthermore, the charging planning module includes a problem modeling unit, an optimization scheduling unit, and an execution management unit; The problem modeling unit receives a planning command or a price fluctuation command and calls the predicted price sequence And determine the global suspension period, query the charging queue table to obtain the rated power, charging lower limit power, charging upper limit power, remaining power, health factor and expected end time of each electric vehicle, and determine the vehicle suspension period, consider the impact of health factor and time period predicted electricity price to set charging power constraints, consider charging demand and charging loss to set power constraints, and consider the upper limit of grid load in the same period Set the total load constraint for each time period, take minimizing the global charging cost as the optimization goal, and build a cost optimization model; The optimization scheduling unit uses Lagrange multipliers to merge charging power constraints, power constraints and total load constraints in the optimization target, rewrites the cost optimization model into an unconstrained optimization problem, and uses the improved rat swarm optimization algorithm to solve and obtain the optimal solution. The Lagrange multipliers in the optimal solution are removed and restored to the optimal charging power sequence group and executed. The optimal charging power sequence group includes the optimal charging power sequence of all electric vehicles. The execution management unit executes the optimal charging power sequence group and clears the optimal charging power sequence of the vehicle that has completed the execution, receives the vehicle clearing command, and queries whether there is an optimal charging power sequence for the vehicle corresponding to the vehicle number in the optimal charging power sequence group. If not, no operation is required. If so, the optimal charging power sequence for the vehicle corresponding to the vehicle number in the optimal charging power sequence group is cleared.

[0012] Furthermore, the construction of the cost optimization model includes the following specific steps: Query the charging queue table to obtain the rated power, lower charging power limit, upper charging power limit, remaining power, health factor and expected end time in each charging record, determine the vehicle suspension period corresponding to each electric vehicle and count the total number of charging periods; The optimization goal is to minimize the global charging cost, that is, the total electricity price of all electric vehicles is the minimum, where the electricity price of each electric vehicle is the predicted electricity price, charging power and duration of each period between the current period and the vehicle stop period. The total electricity price is the sum of the products of all electric vehicles. Set Electric cars in Time period The charging power constraint, Electric cars in Time period Charging power during the period Should be greater than or equal to the lower charging power limit And less than or equal to the upper charging power limit and the theoretical upper limit function The smaller value of , where the theoretical upper limit function Based on health factor Forecast electricity prices by time period Calculate the Theoretical upper limit of charging power of electric vehicles and health factor Proportional to the predicted electricity price during the period Inversely proportional, ; Set The power constraint of an electric vehicle, that is, Electric vehicles during vehicle stop time Remaining power Should be equal to the rated capacity , among which, The amount of charge of an electric vehicle in each period will be based on the health factor There is a proportional loss; Set the total load constraint during the time period, so that the total charging power of all electric vehicles being charged in the same time period is less than or equal to the upper limit of the grid load. ; Integrate optimization objectives, charging power constraints, power constraints, and total load constraints during time periods to build a cost optimization model.

[0013] Furthermore, the optimization scheduling unit uses Lagrange multipliers to rewrite the cost optimization model as an unconstrained optimization problem and uses an improved rat swarm optimization algorithm to plan and generate an optimal charging power sequence group, including the following specific steps: The total number of optimization variables of the explicit cost optimization model is The total number of charging periods for electric vehicles , For the The total number of charging periods for electric vehicles, ; Based on the duality theory, the charging power constraint of a single electric vehicle in a single period is decomposed into three charging power constraint inequalities in the form of less than or equal to, which are respectively multiplied with three different Lagrange multipliers and added to the optimization target to obtain a new optimization target. The total number of optimization variables of the new optimization target is the total number of optimization variables of the cost optimization model. Add 3, and merge all the charging power constraints in the cost optimization model into the original optimization target in the same way to generate the first optimization target. The total number of optimization variables for the first optimization target is ; Since the power constraint of a single electric vehicle is in the form of an equation, the right side of the equal sign is shifted to 0, and the equation is reorganized into a power constraint equation, multiplied by the Lagrange multiplier and added to the first optimization objective. The power constraints of the electric vehicles are incorporated into the first optimization objective to generate the second optimization objective. The total number of optimization variables of the second optimization objective becomes ; The total load constraint of a single period is arranged into a period total load constraint inequality in the form of less than or equal to by shifting terms, and the right side of the inequality sign of the period total load constraint inequality is 0, which is multiplied by the Lagrange multiplier and added to the second optimization objective. The total load constraints of each time period are merged into the second optimization objective to generate an unconstrained optimization problem. The total number of optimization variables of the unconstrained optimization problem becomes , The total number of time periods from the current time period to the global termination time period; The unconstrained optimization problem is solved by improving the rat swarm optimization algorithm, and the construction dimension is The vector space of The mouse with the lowest global electricity cost is determined as the leader mouse in round 0. In each round of iteration, the mouse determined in the previous round of iteration is The servant mice perform random exploration and cross-derivation until the maximum number of iterations is reached to obtain the optimal solution to the unconstrained optimization problem; Extract the optimal solution values, and then reversely restore them to the optimal charging power sequence group according to the arrangement order when defining the solution space, where the remaining The value is the introduced Lagrange multiplier.

[0014] The energy-saving control method of a fast charging station with a peak-valley shifting function is implemented based on an energy-saving control system of a fast charging station with a peak-valley shifting function, and includes the following specific steps: Sensing the first Electric vehicles are connected, and the total number of connected electric vehicles is updated to , get the Electric vehicle charging records And save it in the charging queue list to obtain the previously connected The remaining power of each electric vehicle and update the charging queue list; Perform electricity price forecasting, query the charging records in the charging queue table to determine the global suspension period, and use the time series forecasting network to generate the predicted electricity price sequence ; Execute charging planning, query the charging records of each electric vehicle in the charging queue list to obtain the rated power, charging lower limit power, charging upper limit power, remaining power, health coefficient and expected end time, consider the health coefficient of the electric vehicle and the trend of electricity price changes to build a cost optimization model, introduce Lagrange multipliers to rewrite it as an unconstrained optimization problem, use the improved rat swarm optimization algorithm to solve and generate the optimal solution, remove the Lagrange multipliers in the optimal solution, arrange them in order into the optimal charging power sequence group and execute; In each period before the global suspension period, the prediction error is calculated and compared with the upper limit of the error. If it is greater than the upper limit of the error, the power price prediction is executed. If it is less than or equal to the upper limit of the error, no processing is performed. When a certain electric vehicle is sensed to be disconnected, the corresponding charging record in the charging queue list is deleted, and the optimal charging power sequence group is checked to see whether there is a corresponding vehicle optimal charging power sequence. If so, it is cleared.

[0015] Compared with the prior art, when a new electric vehicle is connected, the present invention obtains charging records through wired connection and wireless link and stores them in the charging queue list, and updates the remaining power of the connected electric vehicles in the charging queue list at the same time; queries the charging queue list to determine the global suspension period, adopts the time series prediction network to analyze the time series evolution of the power grid electricity price to generate a predicted electricity price sequence, and checks the prediction error in each period before the global suspension period to decide whether to regenerate the predicted electricity price sequence; queries the charging queue list to obtain the rated power, charging lower limit power, charging upper limit power, remaining power, health factor and expected end time of the connected electric vehicle, considers the battery life and charging loss caused by the health factor to set the charging power constraint and power constraint respectively, considers the power grid load safety to set the total load constraint of the time period, calculates the global charging cost based on the predicted electricity price sequence and establishes the optimization target, constructs the cost optimization model in combination and rewrites it into an unconstrained optimization problem by introducing Lagrange multipliers, adopts the improved rat swarm optimization algorithm to solve the unconstrained optimization problem, and further extracts and generates the optimal charging power sequence group, so as to realize the adaptive charging planning of electric vehicles that takes into account the charging cost, battery health and power grid safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 Schematic diagram of the energy-saving control system of a fast charging station with peak-valley load shifting function in the present invention; Figure 2 A diagram of a time series prediction network model in the present invention; Figure 3 A flowchart of the pre-training of the time series prediction network using the improved rat colony optimization algorithm in the present invention; Figure 4 This is a flow chart of the energy-saving control method of a fast charging station with peak-valley load shifting function in the present invention. DETAILED DESCRIPTION

[0017] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0018] Example 1

[0019] like Figure 1 As shown, a specific embodiment of the present invention discloses a fast charging station energy-saving control system with peak-valley load shifting function, including an inductive communication module, a power grid monitoring module and a charging planning module; The inductive communication module senses the connection and disconnection of electric vehicles in real time based on the hardware detection circuit. When an electric vehicle is connected, obtain the Electric vehicle charging records And save it in the charging queue list, and simultaneously obtain the connected The remaining power of the electric vehicle is updated and the charging queue list is updated, and then the planning command is broadcast. When an electric vehicle is disconnected, the corresponding charging record in the charging queue list is deleted. and broadcast the vehicle clearance order, Electric vehicle charging records , and They are battery information and charging requirement information respectively. is the vehicle number, , is the total number of ports, which are used to connect charging cables of electric vehicles, including electric two-wheelers and electric three-wheelers; The power grid monitoring module connects with the power grid, receives planning commands, queries the existing charging records in the charging queue table to determine the global suspension period, and uses the time series prediction network to generate the predicted electricity price sequence , check the prediction error in each period before the global suspension period, decide whether to broadcast the electricity price fluctuation command and regenerate the predicted electricity price sequence , where each period lasts for the same length of time, and the predicted electricity price series The length changes dynamically, and the predicted electricity price for all time periods between the current time period and the global termination time period is always recorded; The charging planning module receives the planning command or the electricity price fluctuation command and calls the predicted electricity price sequence , query the charging queue table to obtain the rated power, charging lower limit power, charging upper limit power, remaining power, health coefficient and expected end time of each electric vehicle, comprehensively consider the health coefficient of the electric vehicle and the trend of electricity price changes, build a cost optimization model and rewrite it into an unconstrained optimization problem by introducing Lagrange multipliers, and use the improved rat swarm optimization algorithm to solve the optimal solution of the generated unconstrained optimization problem, further extract and restore it to the optimal charging power sequence group from the current period to the global termination period and execute it, to ensure that each electric vehicle meets the charging demand while minimizing the total electricity cost of all electric vehicles, receive the vehicle clearing command, determine whether there is a corresponding vehicle optimal charging power sequence in the optimal charging power sequence group, and decide whether to clear the corresponding vehicle optimal charging power sequence.

[0020] Further, the inductive communication module includes a port detection unit, a two-way communication unit and a queue storage unit; The port detection unit monitors in real time based on hardware detection circuits The detection current of each port changes to sense the connection of an electric vehicle. When no electric vehicle is connected to the port, the detection current of the port is 0. When an electric vehicle is connected, the connected port forms a loop, and the detection current of the port becomes a non-zero value. When an electric vehicle is disconnected from the connected port, the detection current of the port returns to 0 again. , is the total number of ports; Two-way communication unit When an electric vehicle is connected, the CAN communication protocol is used to communicate with the The battery management system (BMS) of an electric vehicle establishes a wired link and obtains battery information , based on wireless communication technology and The mobile device used by the user of the electric vehicle establishes a wireless link and obtains the charging demand information selected by the user on the mobile device. , combined to generate Electric vehicle charging records At the same time, through the established Before the wired link is reacquired The remaining power of an electric vehicle, including battery information Including rated power , Charging lower limit power , Charging upper limit power , remaining power and health factor , health factor Between 0 and 1, charging demand information Including charging start time and expected end time ; Before the queue storage unit is acquired The remaining power of an electric vehicle, based on the vehicle number Index to the corresponding queue space in the charging queue table and update and replace the recorded remaining power to obtain the first Electric vehicle charging records , according to the vehicle number corresponding to the corresponding queue space stored in the charging queue table, and then broadcast the planning command, when the When an electric vehicle is disconnected, the charging record in the charging queue list Delete and broadcast the vehicle clearing command, which includes the vehicle number. .

[0021] Furthermore, the power grid monitoring module includes a power grid connection unit, an electricity price prediction unit and a time period verification unit; The grid connection unit obtains the grid electricity price in each time period and stores it in the historical time period electricity price table according to the time period sequence; The electricity price prediction unit receives the planning command or correction signal, queries the existing charging records in the charging queue table, takes the time period of the last expected end time in the charging record as the global suspension period, and uses the time series prediction network combined with the historical time period electricity price table for time series reasoning to generate a predicted electricity price sequence. , predicting electricity price series Record the time period forecast electricity prices for all time periods from the current time period to the global suspension time period. If no new planning command is received at the end of the global suspension time period, clear the forecast electricity price sequence; The time period verification unit checks whether there is a predicted electricity price sequence in each time period. If it does not exist, no processing will be done. If it does exist, the historical electricity price table and predicted electricity price sequence will be called. The period electricity price and period predicted electricity price corresponding to the current period are calculated and the prediction error is calculated. When the prediction error is greater than the upper limit of the error, a correction signal is sent to the electricity price prediction unit and an electricity price fluctuation command is broadcast. When the prediction error is less than or equal to the upper limit of the error, the prediction is considered accurate and no processing is performed.

[0022] like Figure 2 As shown, further, the time series prediction network is used to generate the predicted electricity price sequence The specific steps include: Count the total number of time periods from the current time period to the global termination time period The time periods are numbered in chronological order. , where time period number 1 and time period number Corresponding to the current period and the global suspension period respectively; Starting from the current period, search back to the nearest period in terms of time. Historical periods are marked in chronological order. , extracted from the electricity price table for historical periods The electricity price for each historical period , construct the electricity price series for historical periods ; The historical electricity price series is normalized by minimum and maximum Scaling to the same interval to improve the stability and processing efficiency of the time series prediction network, and extracting local features through one-dimensional convolution and ReLU activation function to generate a historical electricity price representation sequence ; The historical electricity price represents the series Capture the historical electricity price representation sequence through the temporal attention layer The single electricity price representation in The time series dependence between the electricity price representations of the time periods is determined according to the strength of the time series dependence. The electricity price representation of each period is assigned Focus on the weight, and use the focus weight to The electricity price representations of the time periods are summed to re-express a single electricity price representation. After all the electricity price representations are re-represented, the historical electricity price representation sequence Reconstruction representation as a weighted electricity price representation sequence ; The weighted electricity price represents the sequence After the gate control loop layer, the gate control loop layer is used to enhance the nonlinear expression ability of the time series prediction network and capture the weighted electricity price representation sequence The memory relationship between adjacent time periods in the process is to determine the retention ratio of the hidden state matrix of a single time period and the weighted electricity price representation of the next time period, and adaptively update the hidden state matrix of the next time period until the hidden state matrix of the historical time period with time period number 0 is obtained. The hidden state matrix of the historical time period with time period number 0 reflects The changing characteristics of electricity price time series during the historical period; The hidden state matrix of the historical period numbered 0 is flattened into the corresponding hidden state vector through the flattening layer, and adjusted to the dimension of The vector is restored to the predicted electricity price sequence from the current period to the global termination period using the inverse minimum and maximum normalization. ,in, and They are the predicted electricity prices for the current period and the global suspension period respectively.

[0023] like Figure 3 As shown, further, the pre-training of the time series prediction network is achieved by improving the rat swarm optimization algorithm, including the following specific steps: Extract the one-dimensional convolution, temporal attention layer and gate control loop layer involved in the time series prediction network hyperparameters and organized into dimensions The hyperparameter vector of is the total number of hyperparameters; Build a dimensional vector space, the hyperparameter vector can be regarded as the position in the vector space, then the pre-training of the time series prediction network is equivalent to Find the optimal position with the best evaluation index in the vector space of ,Since the true value of the pre-training data is known, the evaluation index is defined as the error between the electricity price of the predicted period and the actual period; exist dimensional vector space randomly initialized mice, define this initialization as the 0th iteration, The position of the mouse in iteration 0 is defined as , is the total number of mice, The positions of the mice are used as the time series prediction network Hyperparameters are used to calculate the evaluation index and determine the optimal position with the best evaluation index in the 0th iteration. ; Set the maximum number of iterations , in In the round iteration, , get the The optimal position of the round iteration The corresponding mouse is taken as The leader mouse of the iteration, the rest The servant mouse will Starting from the position of the wheel, move to the optimal position Direction random exploration to get closer to The leader mouse of the iteration, The servant mouse The specific formula for random exploration in round iteration is as follows: ; in, For the The servant mouse The position after random exploration in round iterations, A random number between 0 and 1, reflecting the The exploration randomness of the servant mouse, For the The exploration step size of the round iteration increases with the number of iterations. The increase of Wandering around, The exploration step size of the round iteration The specific formula is as follows: ; in, is a random number between 1 and 2; To prevent The swarm of mice consisting of 10 mice falls into the local optimal problem during exploration. For example, the initialization When the positions of the mice are determined, a mouse happens to be near the local optimal position and the evaluation indicators of the remaining mice are weaker than those of the mice near the local optimal position. As a result, all mice explore the local optimal position and gradually deviate from the global optimal position during the iterative update process. Therefore, cross-derivation is introduced to improve the mouse swarm optimization algorithm. The servant mouse After random exploration in rounds of iterations, multiple groups of non-repeating servant pairs are randomly selected. For each group of servant pairs, the two servants are randomly linearly combined in position or the values ​​of different dimensions in the two positions are randomly exchanged to derive servants in new positions and randomly replace any servant in the servant pair. After the cross-derivation is completed, Round Iteration The positions of the mice are used as the time series prediction network hyperparameters to calculate the evaluation index and determine the The optimal position with the best round of iterative evaluation indicators ; Continue to start the next round of iteration and perform random exploration and cross-derivation in sequence until the maximum number of iterations is reached ,at this time All mice gathered at the optimal location , the best position In The hyperparameters are the hyperparameters determined when the time series prediction network pre-training is completed.

[0024] Furthermore, the charging planning module includes a problem modeling unit, an optimization scheduling unit, and an execution management unit; The problem modeling unit receives a planning command or a price fluctuation command and calls the predicted price sequence And determine the global suspension period, query the charging queue table to obtain the remaining power, health factor and expected end time of each electric vehicle, and use the period of expected end time as the corresponding vehicle suspension period. Consider the indirect impact of the health factor of each electric vehicle battery and the predicted electricity price during the period to set the charging power constraint, consider the charging demand of each electric vehicle and the charging loss of the battery to set the power constraint, and consider the upper limit of the grid load in the same period. Set the total load constraint for each time period, take minimizing the global charging cost as the optimization goal, and build a cost optimization model; The optimization scheduling unit uses Lagrange multipliers to merge the charging power constraint, power constraint and total load constraint of the time period into the optimization target, so that the cost optimization model is rewritten as an unconstrained optimization problem. The improved rat swarm optimization algorithm is used to solve the unconstrained optimization problem to generate the optimal solution. The Lagrange multipliers introduced in the optimal solution are removed and arranged in sequence as the optimal charging power sequence group from the current time period to the global suspension period, so that each electric vehicle can complete the charging task before the corresponding vehicle suspension period and the total electricity cost of all electric vehicles to complete charging is the lowest. Among them, the optimal charging power sequence group includes the vehicle optimal charging power sequence of all electric vehicles, and the vehicle optimal charging power sequence records the optimal charging power of each time period between the current time period and the corresponding vehicle suspension period. The execution management unit executes the optimal charging power sequence group, controls the charging power of each electric vehicle in each time period, and when an electric vehicle executes to the corresponding vehicle suspension time period, clears the corresponding vehicle optimal charging power sequence in the optimal charging power sequence group, receives the vehicle clearing command, and queries whether there is a vehicle optimal charging power sequence corresponding to the vehicle number in the optimal charging power sequence group. If not, it indicates that the battery disconnection of the electric vehicle is a normal behavior after charging is completed, and no operation is required. If so, it indicates that the battery disconnection of the electric vehicle is an abnormal behavior before charging is completed, and clears the vehicle optimal charging power sequence corresponding to the vehicle number in the optimal charging power sequence group and which has not been executed.

[0025] Furthermore, the construction of the cost optimization model includes the following specific steps: Query the charging queue list to obtain the existing Charging records , is the total number of ports. Electric vehicle charging records , get the rated power , Charging lower limit power , Charging upper limit power , remaining power , Health Factor and expected end time , because when a new electric vehicle is connected, the inductive communication module will update the remaining power of the connected electric vehicle, so the The remaining power of an electric vehicle In fact, the current period Real-time remaining power , based on the expected end time Determine The vehicle stop time of each electric vehicle is counted based on the current time period and the vehicle stop time period. Total number of charging sessions for electric vehicles , then The vehicle stop period of an electric vehicle can be defined as , , is the total number of ports; The optimization goal is to minimize the global charging cost, that is, the total electricity price of all electric vehicles from the current period to the global suspension period is the least, and the electricity price of each electric vehicle in a single period is equal to the product of the predicted electricity price of the single period and the charging power of the single period. The optimization goals are as follows: , in, For the Time period The electricity price can be predicted from the time period Get For the Electric cars in Time period The charging power during the period, The duration of a fixed period is the product of the duration of the period and the charging power of the period, which is exactly the charging power of the period; No. Electric cars in Time period Charging power during the period Should always meet or be equal to the lower charging power limit And less than or equal to the upper charging power limit , considering the health factor The impact of health factors The lower battery receives a higher charging power during the period The battery life is seriously damaged, and the electricity price is predicted based on the time period. The impact of When it is higher, the charging power of the time period We also hope to be lower, set the charging power constraint as follows: , in, is the theoretical upper limit function, used to calculate the health coefficient Forecast electricity prices by time period Determine the time The theoretical upper limit power that an electric vehicle can accept for charging, the theoretical upper limit function The specific formula is as follows: , in, represents the average electricity price of the power grid, which is given by the theoretical upper limit function It can be seen that when the health factor Very low and time period forecast electricity price When it is very high, The theoretical upper charging power limit of an electric vehicle will be infinitely close to the lower charging power limit. , when the health factor Very high and time period forecast electricity price When the electricity price is far below the average, The theoretical upper limit of charging power for electric vehicles can even be greater than the upper limit of charging power. ; No. Electric vehicles must meet charging requirements, that is, vehicle stop time Remaining power Should be equal to the rated capacity , considering the The battery of an electric vehicle has charging loss in every period, and the charging loss is related to the health factor. Equivalently, set the power constraints as follows: , in, For the Electric cars in Time period The remaining power; The total charging power of all electric vehicles being charged in the same period cannot exceed the upper limit of the grid load. To reduce the risk of accidents, set the total load constraints for the time period. Time period For example, the details are as follows: , in, Indicates the current period Access Electric vehicles in the Time period The set of vehicle numbers of electric vehicles that are still connected, Always a collection A subset of Integrate optimization objectives, charging power constraints, power constraints, and total load constraints during time periods to build a cost optimization model.

[0026] Furthermore, the optimization scheduling unit uses Lagrange multipliers to rewrite the cost optimization model as an unconstrained optimization problem and uses an improved rat swarm optimization algorithm to plan and generate an optimal charging power sequence group, including the following specific steps: The optimization variables of the cost optimization model are the charging power of each electric vehicle in each period. The total number of charging periods for electric vehicles is , then the total number of optimization variables of the cost optimization model is The total number of charging periods for electric vehicles ; Get optimization goals ,in, For the Time period Forecast electricity prices for the time period For the Electric cars in Time period The charging power during the period, The duration of a fixed period of time; Get the Electric cars in Time period Charging power during the period The charging power constraints are as follows: , in, is the theoretical upper limit function, , and They are the lower limit power, upper limit power and health factor. The charging power constraint can be decomposed into three charging power constraint inequalities, as follows: , , , Three Lagrange multipliers are introduced to multiply the three charging power constraint inequalities respectively, and then added to the optimization target to obtain a new optimization target. Based on the duality theory, the new optimization target is equivalent to the combination of the original optimization target and the charging power constraint. However, the optimization variables of the new optimization target are added with the three newly introduced Lagrange multipliers on the basis of the original optimization variables. All the charging power constraints in the original cost optimization model are merged into the original optimization target to become the first optimization target. The total number of optimization variables of the first optimization target is equal to 4 times the sum of the total number of charging periods. This is because the total number of optimization variables in the original cost optimization model is equal to the total number of charging periods. Each electric vehicle has a charging power constraint in each period before the vehicle stop period. The charging power constraint in each charging period corresponds to three Lagrange multipliers. After merging, the total The total number of optimization variables; For The power constraints of electric vehicles are as follows: , in, and The current period , No. Time period and vehicle stop period The remaining power, is the rated capacity, The power constraint of an electric vehicle is summarized as the power constraint equation, as follows: , Introduce the Lagrange multiplier to multiply the power constraint equation and add it to the first optimization objective. The power constraints of the electric vehicles are merged into the first optimization objective to generate the second optimization objective. The total number of optimization variables of the second optimization objective becomes ; No. Time period The total load constraints for the time period are as follows: , in, Indicates the current period Access Electric vehicles in the Time period The set of vehicle numbers of the electric vehicles that are still connected is organized into a time period total load constraint inequality in the form of less than or equal to, as follows: , Introduce the Lagrange multiplier to multiply the total load constraint inequality of the time period and add it to the second optimization objective. It is known that the total number of time periods from the current time period to the global termination period is , so the total load constraint for the time period is indivual, , the global suspension period is The latest vehicle stop period among the electric vehicles will be The total load constraints of each time period are merged into the second optimization objective to generate an unconstrained optimization problem. The total number of optimization variables of the unconstrained optimization problem becomes ; For unconstrained optimization problems, we can solve them by improving the rat swarm optimization algorithm and construct a The solution of the unconstrained optimization problem can be regarded as the position in the vector space. The evaluation index is defined as the global electricity cost, which is equivalent to dimensional vector space to find an optimal position with the lowest global electricity cost, using the same idea as the pre-training of the time series prediction network, and initialize The mouse with the lowest cost is determined as the leader mouse in the 0th iteration. In each subsequent iteration, the mouse determined in the previous iteration is The servant mice conduct random exploration and cross-derivation towards the leader mouse until the maximum number of iterations is reached to obtain the optimal solution to the unconstrained optimization problem; Extract the optimal solution values, corresponding to The optimal charging power of each electric vehicle before its respective vehicle stop period is reversely restored to the optimal charging power sequence group according to the arrangement order when defining the solution space.

[0027] Example 2

[0028] like Figure 4As shown, the energy-saving control method of a fast charging station with a peak-valley shifting function is implemented based on an energy-saving control system of a fast charging station with a peak-valley shifting function, and includes the following specific steps: Based on hardware detection circuit sensing The number of connected electric vehicles is updated to , obtain the first Electric vehicle charging records And save it in the charging queue list to obtain the previously connected The remaining power of each electric vehicle and update the charging queue list; Perform electricity price forecasting, query the existing charging records in the charging queue table to determine the global suspension period, and use the time series forecasting network to generate the predicted electricity price sequence from the current period to the global suspension period ; Execute charging planning, query the charging records of each electric vehicle in the charging queue list to obtain the rated power, charging lower limit power, charging upper limit power, remaining power, health coefficient and expected end time, comprehensively consider the health coefficient of the electric vehicle and the trend of electricity price changes to build a cost optimization model, introduce Lagrange multipliers to rewrite it as an unconstrained optimization problem, use the improved rat swarm optimization algorithm to solve and generate the optimal solution, eliminate the Lagrange multipliers in the optimal solution, arrange them in order into the optimal charging power sequence group and execute; The prediction error is calculated in each period before the global termination period and compared with the upper limit of the error. If the prediction error is greater than the upper limit of the error, the power price prediction is executed. If the prediction error is less than or equal to the upper limit of the error, no processing is performed. Based on the hardware detection circuit, a disconnection of a certain electric vehicle is sensed, the corresponding charging record in the charging queue list is deleted, and it is determined whether there is an optimal charging power sequence for the vehicle corresponding to the disconnected electric vehicle in the optimal charging power sequence group; If it does not exist, no processing is performed. If it exists, the vehicle optimal charging power sequence corresponding to the disconnected electric vehicle is cleared.

[0029] The present invention discloses an energy-saving control system and method for a fast charging station with a peak-valley load shifting function, comprising an inductive communication module, a power grid monitoring module and a charging planning module; the inductive communication module senses the access of electric vehicles, stores corresponding charging records and updates the remaining power of other connected electric vehicles; the power grid monitoring module determines a global suspension period based on the charging records, generates a predicted electricity price sequence using a time series prediction network, and checks whether the prediction error decision of each period regenerates the predicted electricity price sequence during the charging process; the charging planning module calls the predicted electricity price sequence, obtains the charging records of connected electric vehicles, considers battery health, electricity price changes and power grid load safety setting constraints, takes minimizing the global electricity cost as the goal, constructs a cost optimization model in combination and introduces Lagrange multipliers to rewrite it into an unconstrained optimization problem, uses an improved rat swarm optimization algorithm to solve and extract and restore it into an optimal charging power sequence group, and realizes adaptive charging planning that takes into account cost, safety and battery health.

[0030] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A fast charging station energy-saving control system with peak-valley load shifting function, characterized in that: It includes inductive communication module, power grid monitoring module and charging planning module; When the inductive communication module senses that an electric vehicle is connected, the corresponding charging record is stored in the charging queue table, the remaining power of other connected electric vehicles is updated, and a planning command is broadcast; The power grid monitoring module receives the planning command, determines the global suspension period based on all charging records in the charging queue table, generates a predicted electricity price sequence using a time series prediction network in combination with the electricity price of the historical period, checks the prediction error in each period of the charging process, decides whether to regenerate the predicted electricity price sequence and broadcasts the electricity price fluctuation command; The charging planning module receives a planning command or an electricity price fluctuation command, calls a predicted electricity price sequence, queries a charging queue table to obtain charging records of connected electric vehicles, considers the health coefficient of the electric vehicle battery, the electricity price change trend and the grid load safety setting constraints in each period, takes minimizing the global electricity cost as the optimization goal, and constructs a cost optimization model in combination. Lagrange multipliers are introduced to rewrite the cost optimization model as an unconstrained optimization problem and an improved rat swarm optimization algorithm is used to obtain the optimal solution. The Lagrange multipliers in the optimal solution are eliminated and restored to the optimal charging power sequence group in sequence.

2. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1, characterized in that: The charging planning module includes a problem modeling unit and an optimization scheduling unit; The problem modeling unit receives a planning command or an electricity price fluctuation command, calls a predicted electricity price sequence, queries a charging queue table to obtain a charging record of each electric vehicle, wherein the charging record includes a remaining power, a health factor, and an expected end time, determines a global suspension period and a vehicle suspension period of each electric vehicle, sets a charging power constraint considering the influence of the health factor and the predicted electricity price during the period, sets a power constraint considering the charging demand and the charging loss, sets a total load constraint during the period considering the upper limit of the power grid load during the same period, takes minimizing the global charging cost as an optimization goal, and constructs a cost optimization model; The optimization scheduling unit uses Lagrange multipliers to merge the charging power constraints, power constraints and total load constraints of the time period into the optimization objectives. Based on the duality theory, Lagrange multipliers are introduced to rewrite the cost optimization model into an unconstrained optimization problem. The improved rat swarm optimization algorithm is used to solve the optimal solution of the generated unconstrained optimization problem. The Lagrange multipliers in the optimal solution are eliminated and restored to the optimal charging power sequence group in sequence.

3. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1 or 2, characterized in that: The improved rat swarm optimization algorithm comprises the following specific steps: Determine the dimension of the vector space and define the evaluation index, and randomly initialize it in the vector space in the 0th iteration A mouse, The evaluation index is calculated for each mouse position to determine the optimal position for round 0. is the total number of mice; Set the maximum number of iterations , in In the round of iteration, based on The optimal position of the wheel determines the The leader mouse of the iteration, each of the other servant mice starts from the Starting from the position of the wheel, The optimal position direction of the round is randomly explored, and the distance of random exploration increases with the number of iterations. decreases with the increase of ; Randomly select multiple groups of non-repetitive servant mouse pairs, cross-derive the two servant mice in each group of servant mouse pairs, derive the servant mouse at the new position and randomly replace one servant mouse in the servant mouse pair through the linear combination of the positions of the two servant mice or the random exchange of elements of different dimensions in the positions; Based on In round iteration The evaluation index is recalculated based on the position of the mouse and the The optimal position of the round, start the next round of iteration and perform random exploration and cross-derivation until the maximum number of iterations is reached ,at this time The location where the mice gather is the optimal location.

4. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1 or 2, characterized in that: The construction of the cost optimization model includes the following specific steps: Query the charging queue table to obtain the rated power, lower charging power limit, upper charging power limit, remaining power, health factor and expected end time of each electric vehicle charging record, determine the corresponding vehicle suspension period and count the total number of charging periods; The optimization goal is to minimize the global charging cost, that is, the sum of the electricity prices of all connected electric vehicles is minimized, where the electricity price of a single electric vehicle is the sum of the product of the predicted electricity price of each period between the current period and the vehicle stop period, the charging power of the period, and the duration of the period; Set the charging power constraint of a single electric vehicle in each time period. The charging power of a single electric vehicle in each time period is greater than or equal to the lower charging limit power and less than or equal to the smaller value of the upper charging limit power and the theoretical upper limit function. The theoretical upper limit function is proportional to the health coefficient and the inverse of the predicted electricity price during the time period, and is used to calculate the theoretical upper charging limit power of a single electric vehicle. Set the power constraint of a single electric vehicle and the total load constraint of a time period, that is, the remaining power of a single electric vehicle during the vehicle suspension period is equal to the rated power and the total charging power of all connected electric vehicles in the same time period is less than or equal to the upper limit of the grid load. The charging capacity of a single electric vehicle in each time period will be proportionally lost according to the health factor; Integrate optimization objectives, charging power constraints, power constraints, and total load constraints during time periods to build a cost optimization model.

5. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1 or 2, characterized in that: The introduction of Lagrange multipliers to rewrite the cost optimization model as an unconstrained optimization problem includes the following specific steps: The total number of optimization variables in the cost optimization model is the sum of the total number of charging periods for connected electric vehicles. ; Based on the duality theory, the charging power constraint of each electric vehicle in each period is decomposed into three charging power constraint inequalities in the form of less than or equal to, which are multiplied by three Lagrange multipliers respectively and added to the optimization target to generate the first optimization target. The total number of optimization variables for the first optimization target is ; The power constraint of each electric vehicle is transformed into a power constraint equation by shifting terms, multiplied by the Lagrange multiplier and added to the first optimization objective to generate the second optimization objective. The total number of optimization variables for the second optimization objective becomes , is the total number of connected electric vehicles; By shifting the terms, the total load constraint of each time period is organized into a total load constraint inequality of the form less than or equal to, and the right side of the inequality sign of the total load constraint inequality is 0. It is multiplied by the Lagrange multiplier and added to the second optimization objective to generate an unconstrained optimization problem. The total number of optimization variables of the unconstrained optimization problem is , The total number of time periods from the current time period to the global termination time period.

6. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1, characterized in that: The power grid monitoring module includes a power grid connection unit, an electricity price prediction unit and a time period verification unit; The grid connection unit obtains the electricity price in each time period and stores it in the historical time period electricity price table; The electricity price prediction unit receives a planning command or a correction signal, queries the charging queue table, takes the time period of the last expected end time in the existing charging record as the global suspension period, combines the historical time period electricity price table, and uses the time series prediction network to generate a predicted electricity price sequence. If no new planning command is received at the end of the global suspension period, the predicted electricity price sequence is cleared; The time period verification unit queries whether there is a predicted electricity price sequence in each time period. If not, no processing is performed. If so, the time period electricity price and the time period predicted electricity price of the current time period are called and the prediction error is calculated. The prediction error is compared with the error upper limit to decide whether to send a correction signal to the electricity price prediction unit and broadcast an electricity price fluctuation command.

7. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1 or 6, characterized in that: The method of using a time series forecasting network to generate a forecasted electricity price sequence includes the following specific steps: Count the total number of time periods from the current time period to the global termination time period And mark the time period numbers in sequence; Extract the electricity price closest to the current period from the historical period electricity price table The electricity prices of historical periods are used to construct the electricity price series of historical periods; Normalize the electricity price series of the historical period and extract features through one-dimensional convolution and ReLU activation function to generate a historical electricity price representation series; The temporal attention layer captures the relationship between a single electricity price representation and the rest of the historical electricity price representation sequence The time dependence of the electricity price representation is adaptively assigned to the rest The different attention weights of the electricity price representations are weighted to re-represent a single electricity price representation. Each electricity price representation is re-represented to obtain a weighted electricity price representation sequence; The gate-controlled recurrent layer is used to capture the retention ratio of the hidden state matrix of a single period in the weighted electricity price representation sequence and the weighted electricity price representation of the next period, and the hidden state matrix of the next period is updated and generated, and the hidden state matrix of the historical period with period number 0 is obtained by loop execution; The hidden state matrix of the historical period numbered 0 is flattened by a flattening layer, dimension reduction is performed through a fully connected layer, and anti-minimum-maximum normalization is used to restore it to the predicted electricity price series.

8. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1, characterized in that: The inductive communication module includes a port detection unit, a two-way communication unit and a queue storage unit; The port detection unit monitors the detection current of all ports based on the hardware detection circuit. When the detection current of the port changes from 0 to non-0, it senses the connection of the electric vehicle. When the detection current of the port changes from non-0 to 0, it senses the disconnection of the electric vehicle. The two-way communication unit establishes a wired link with the battery management system of the newly connected electric vehicle based on the CAN communication protocol to obtain battery information, establishes a wireless link with the mobile device of the user of the newly connected electric vehicle based on the wireless communication technology to obtain charging demand information, combines and generates charging records, and re-acquires the remaining power of the previously connected electric vehicle through the established wired link; The queue storage unit obtains the remaining power of the previously connected electric vehicle and replaces it accordingly in the charging queue list, obtains the charging record of the newly connected electric vehicle and stores it in the charging queue list, broadcasts the planning command, and when an electric vehicle is disconnected, deletes the corresponding charging record and broadcasts the vehicle clearing command.

9. The fast charging station energy-saving control system with peak-valley shifting function as claimed in claim 1, characterized in that: The charging planning module also includes an execution management unit; The execution management unit executes the optimal charging power sequence group and clears the vehicle optimal charging power sequence that has been executed, receives a vehicle clearing command, queries whether there is a vehicle optimal charging power sequence corresponding to the vehicle number in the optimal charging power sequence group, if not, no operation is required, if yes, clears the vehicle optimal charging power sequence corresponding to the vehicle number in the optimal charging power sequence group, wherein the vehicle clearing command is issued when the induction communication module senses that the electric vehicle is disconnected.

10. A fast charging station energy-saving control method with peak-valley load shifting function, characterized in that: The specific steps include: Sensing the first Electric vehicles have been connected, and the total number of connected electric vehicles has been updated to , get the The charging records of electric vehicles are stored in the charging queue list, and the previous The remaining power of each electric vehicle and update the charging queue list; Perform electricity price prediction, query the charging records in the charging queue table to determine the global suspension period, and use the time series prediction network to generate the predicted electricity price sequence; Execute charging planning, query the charging records of each electric vehicle in the charging queue to obtain the rated power, charging lower limit power, charging upper limit power, remaining power, health factor and expected end time, consider the health factor of the electric vehicle battery, the trend of electricity price changes and the safety setting constraints of the grid load in each period, take minimizing the global electricity cost as the optimization goal, combine and construct the cost optimization model, introduce Lagrange multipliers to rewrite the cost optimization model as an unconstrained optimization problem, and use the improved rat swarm optimization algorithm to obtain the optimal solution, remove the Lagrange multipliers in the optimal solution, and restore it to the optimal charging power sequence group in sequence and execute it; In each period before the global suspension period, the prediction error is calculated and compared with the upper limit of the error. If it is greater than the upper limit of the error, the power price prediction is executed. If it is less than or equal to the upper limit of the error, no processing is performed. When it is sensed that an electric vehicle is disconnected, the corresponding charging record in the charging queue list is deleted, and the optimal charging power sequence group is checked to see whether there is a corresponding vehicle optimal charging power sequence. If so, it is cleared.

Images