Electric two-wheeler charging system, method and storage medium based on dynamic allocation and adjustment

Abstract

The application discloses an electric two-wheeled vehicle charging system and method based on dynamic allocation and adjustment, and a storage medium, relates to the electric two-wheeled vehicle charging management technical field, and the system and the method realize digital modeling and dynamic correlation analysis of electric two-wheeled vehicle charging pile cluster by building a virtual cluster system, rely on multi-dimensional data integration of geographical distribution, capacity scale and power consumption attribute, combine the comprehensive computer mechanism of space correlation and load correlation, form the digital twin model that can map the physical cluster state in real time, through the deep learning of the historical charging demand time sequence characteristics of the LSTM model, combine the quantitative analysis of the dynamic connection index and the pre-connection reasonable value, build the smooth transition mechanism of the current strategy and the prediction strategy, ensure that the current charging allocation and adjustment strategy not only meets the current demand, but also maximizes the convenience of subsequent strategy connection and switching.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric two-wheeler charging management, more particularly, it relates to an electric two-wheeler charging system and method based on dynamic allocation and adjustment and a storage medium. BACKGROUND

[0002] With the popularity of electric two-wheelers in urban micro-mobility scenarios, the contradiction between their fragmented and high-frequency charging needs and grid load management and efficient use of charging resources has become increasingly prominent. Existing electric two-wheeler charging management systems mostly use static scheduling mode, lacking dynamic modeling and collaborative scheduling capabilities for charging pile clusters: on the one hand, traditional solutions only allocate charging based on the real-time state of a single cluster, without considering multi-dimensional features such as geographic distribution and electricity consumption attributes, resulting in low resource allocation efficiency between adjacent clusters. Community test data shows that the average idle rate of charging piles is as high as 35%; on the other hand, existing technologies mostly rely on rule engines or simple heuristic algorithms to generate scheduling strategies, which cannot effectively handle fluctuations in peak and valley electricity prices for residential electricity consumption.

[0003] In the field of demand prediction and strategy connection, existing solutions have significant technical bottlenecks: statistical analysis-based prediction models are difficult to capture the nonlinear fluctuations of electric two-wheeler charging demand (such as demand spikes during food delivery peak hours), and the 1-hour prediction error rate of traditional ARIMA models in a certain commercial district test exceeds 30%; at the same time, strategy generation does not consider the smoothness of the connection between the previous and subsequent periods, and when the grid load or user demand changes, strategy switching often leads to sudden increases and decreases in charging pile power, with a power fluctuation amplitude of up to 40% in a certain case, severely affecting grid safety and equipment life. In addition, existing systems lack optimization of electric two-wheeler-specific features (such as the proportion of emergency needs with SOC <20% and super-fast charging power distribution), and cannot meet the needs of high-time-efficiency users such as riders, with an average waiting time of more than 15 minutes.

[0004] To address the above technical problems, the present application provides an electric two-wheeler charging system and method based on dynamic allocation and adjustment and a storage medium. SUMMARY

[0005] To address the deficiencies in the prior art, the present application aims to provide an electric two-wheeler charging system and method based on dynamic allocation and adjustment and a storage medium.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0007] The electric two-wheeler charging system based on dynamic allocation and adjustment comprises a virtual cluster establishment unit for determining all charging pile clusters contained in a charging area and building a virtual cluster system;

[0008] The current feasible strategy analysis unit determines the current feasible strategy according to the periodic virtual cluster system;

[0009] The predicted feasible strategy analysis unit determines the predicted charging demand characteristics of each charging pile cluster, and further determines the predicted feasible strategy.

[0010] The charging distribution adjustment determination unit generates a plurality of strategy connection combinations according to the current feasible strategy and the predicted feasible strategy, determines the dynamic connection index of each strategy connection combination, further determines the pre-connection reasonable value of each current feasible strategy, and selects the current determination strategy from the current feasible strategy, and uses the current determination strategy to distribute and adjust the charging of all electric two-wheeled vehicles in the charging area.

[0011] Further, the current feasible strategy is determined according to the periodic virtual cluster system, specifically as follows: the actual charging demand characteristics of each charging pile cluster are determined, the actual charging demand characteristics of each charging pile cluster are converted into input parameters and introduced into the virtual cluster system, the virtual cluster system generates a plurality of charging distribution adjustment strategies, and each generated charging distribution adjustment strategy is marked as a current feasible strategy.

[0012] Further, the predicted feasible strategy is determined according to the predicted charging demand characteristics of each charging pile cluster, and the predicted feasible strategy is further determined, specifically as follows: the predicted charging demand characteristics of each charging pile cluster are converted into input parameters and introduced into the virtual cluster system, the virtual cluster system generates a plurality of charging distribution adjustment strategies, and each generated charging distribution adjustment strategy is marked as a predicted feasible strategy.

[0013] Further, the predicted charging demand characteristics of the charging pile cluster are determined in the following manner: a charging pile cluster is obtained in the previous continuous determination of a plurality of actual charging demand characteristics, and the plurality of actual charging demand characteristics are integrated into an actual charging demand time sequence characteristic set in a time sequence set manner, the actual charging demand time sequence characteristic set is introduced into the cluster charging demand prediction model corresponding to the charging pile cluster, and the cluster charging demand prediction model exports the predicted charging demand characteristics of the charging pile cluster.

[0014] Further, a plurality of strategy connection combinations are generated according to the current feasible strategy and the predicted feasible strategy, specifically as follows: each current feasible strategy is matched with each predicted feasible strategy one by one, and the matched current feasible strategy and predicted feasible strategy are marked as a strategy connection combination.

[0015] Further, the current feasible strategy with the largest pre-connection reasonable value is marked as the current determination strategy.

[0016] Further, the determination method of the dynamic connection index of the strategy connection combination is as follows: the feature set of the current feasible strategy in the strategy connection combination is determined , determine the feature set of the predicted feasible strategy in the strategy connection combination , i represents the i-th feature in the current feasible strategy or the predicted feasible strategy, the similarity of the two feature sets is calculated by using the cosine similarity formula, and the calculated similarity is the dynamic connection index of the strategy connection combination.

[0017] Further, the determination method of the pre-connection reasonable value of the current feasible strategy is as follows: a current feasible strategy is selected, all strategy connection combinations containing the current feasible strategy are marked as possible connection combinations, the dynamic connection indexes of all possible connection combinations are calculated, the average dynamic connection index is calculated, all possible connection combinations are matched two by two, the dynamic connection indexes of the two matched possible connection combinations are calculated, the dynamic connection difference index is calculated, the average dynamic connection difference index is calculated, and the average dynamic connection index is compared with the average dynamic connection difference index, and the pre-connection reasonable value of the current feasible strategy is calculated.

[0018] Further, the electric two-wheeled vehicle charging method based on dynamic allocation and adjustment has the following steps:

[0019] Step one: build a virtual cluster system;

[0020] Step two: periodically determine the current feasible strategy and the predicted feasible strategy in the virtual cluster system;

[0021] Step three: select a current determination strategy from the current feasible strategy, and use the current determination strategy to allocate and adjust the charging of all electric two-wheeled vehicle charging piles in the charging area.

[0022] Compared with the prior art, the present application has the following advantages:

[0023] The system and method of the present application realize digital modeling and dynamic correlation analysis of the electric two-wheeled vehicle charging pile cluster by building a virtual cluster system, rely on multi-dimensional data integration of geographic distribution, capacity scale and power consumption attributes, combine spatial correlation and load correlation comprehensive computer mechanism, form a digital twin model that can real-time map the physical cluster state, through deep learning of the historical charging demand time sequence characteristics by the LSTM model, combined with the quantitative analysis of the dynamic connection index and the pre-connection reasonable value, a smooth transition mechanism of the current strategy and the predicted strategy is constructed, which ensures that the current charging allocation and adjustment strategy not only meets the current demand, but also maximizes the convenience of the connection and switching of the subsequent strategy, through the closed-loop control of "real-time demand analysis-future trend prediction-strategy connection evaluation", the intelligent upgrading of the urban micro-mobility charging network is provided with core technical support. BRIEF DESCRIPTION OF DRAWINGS

[0024] Fig. 1 Flow chart of the operation of the dynamically allocated and regulated electric two-wheeler charging system;

[0025] Fig. 2 Flow chart of the determination of the pre-connection reasonable value of the current feasible strategy. DETAILED DESCRIPTION

[0026] Embodiment one: reference Figs. 1-2 , based on the dynamically allocated and regulated electric two-wheeler charging system, including virtual cluster establishment unit, current feasible strategy analysis unit, predicted feasible strategy analysis unit, charging allocation and regulation determination unit.

[0027] The virtual cluster establishment unit determines all charging pile clusters contained in the charging area, and builds a virtual cluster system, the building steps are: S1: determine the geographic distribution information (the geographic distribution information is described by geographic coordinates and other information), the capacity scale (the capacity scale is the number of charging piles contained in each charging pile cluster) and the electricity attribute (usually residential electricity, commercial electricity, etc., the characteristics of residential electricity are that the electricity price fluctuates with time, the off-peak electricity price is cheaper, for example: a charging pile cluster accesses residential electricity, the peak period (7:00-22:00) electricity price 0.55 yuan / kWh, the valley period (22:00-7:00) electricity price 0.3 yuan / kWh, the characteristics of commercial electricity are that the electricity price is stable, for example: a charging pile cluster accesses commercial electricity, the electricity price is 0.8 yuan / kWh all day long, and there is no time period distinction);

[0028] S2: design the virtual cluster system architecture, and abstract modeling for the charging pile cluster: charging pile cluster object definition: class PhysicalCluster:

[0029] def __init__(self, cluster_id, coords, capacity, tariff_type,

[0030] pile_types, service_scenario):

[0031] self.id = cluster_id # cluster ID

[0032] self.coordinates = coords # geographic coordinates

[0033] self.pile_count = capacity # number of charging piles

[0034] self.tariff_type = tariff_type # electricity price type

[0035] self.pile_types = pile_types # pile type proportion (fast charging / slow charging)

[0036] self.scenario = service_scenario # service scenario (community / business district, etc.)

[0037] Association relationship construction: spatial association: construct adjacency matrix based on road network distance (threshold < 3 km); load association: calculate Pearson correlation coefficient (threshold > 0.7); comprehensive association degree: R = 0.6 x spatial association + 0.4 x load association (R > 0.6 to form a virtual cluster), construct a digital twin model: C(t) = f(P(t), S(t), G(t), H), where: C(t) is the cluster state vector at time t, P(t) is the charging pile state matrix, S(t) is the vehicle demand vector, G(t) is the power grid state vector, and H is the historical feature matrix; allocation strategy engine: multi-objective optimization algorithm: NSGA-III (optimization objectives: load balancing, user waiting time, charging cost), NSGA-III algorithm characteristics: can generate a set of non-dominated solutions (usually 20-50), representing different trade-offs of optimization objectives; example: in a certain test, 32 Pareto optimal solutions were generated for 3 optimization objectives; reinforcement learning model: DDPG (dynamic adjustment of allocation strategy, reward function includes power grid coordination contribution), through the exploration-exploitation mechanism, different strategy variants can be generated, and the power grid coordination contribution in the reward function can guide the generation of diversified strategies, and a virtual cluster system is built.

[0038] The current feasible strategy analysis unit periodically determines the actual charging demand characteristics of each charging pile cluster (the time length of the corresponding period interval is adjusted according to the charging demand of the charging pile cluster, and each charging pile cluster corresponds to an actual charging demand characteristic. The actual charging demand characteristic contains the number of vehicles to be charged (data source: charging pile operator platform real-time state (idle / occupied), user APP charging request queue; collection method: through MQTT protocol from the charging pile controller, or from the user end APP request interface to pull; example data: a cluster currently shows 3 vehicles to be charged, data from charging pile real-time state reporting), emergency charging demand proportion (data source: vehicle battery SOC (State of Charge) data (externally collected); calculation logic: emergency demand is defined as the proportion of vehicles with SOC <20% (internal rule); data flow: vehicle-mounted BMS transmits SOC through CAN bus→charging pile interface collection→edge node calculation proportion), demand power distribution (data source: user-selected charging mode (fast charging / slow charging), vehicle-supported maximum charging power; collection method: user APP selection record (such as "quick charging" button click), vehicle parameter interface acquisition; typical data: 8 fast charging demands and 5 slow charging demands in a cluster), adjacent cluster overflow demand (data source: number of vehicles to be charged and charging pile utilization rate of adjacent clusters (externally collected); calculation logic: when the utilization rate of the adjacent cluster is >80%, overflow demand = the number of vehicles to be charged in the cluster x 20% (internal algorithm); data link: cloud obtains the state of adjacent clusters→calculates overflow based on spatial correlation→returns to the current cluster), etc.), converts the actual charging demand characteristics of each charging pile cluster into input parameters and imports them into the virtual cluster system, and the virtual cluster system generates multiple charging distribution adjustment strategies (each charging distribution adjustment strategy indirectly affects power distribution by controlling the charging pile power of the charging pile cluster, and each charging distribution adjustment strategy is a Pareto optimal solution). Each generated charging distribution adjustment strategy is labeled as a current feasible strategy.

[0039] The predicted feasible strategy analysis unit synchronously determines the predicted charging demand characteristics of each charging pile cluster, converts the predicted charging demand characteristics of each charging pile cluster into input parameters and imports them into the virtual cluster system, and the virtual cluster system generates multiple charging distribution adjustment strategies. Each generated charging distribution adjustment strategy is labeled as a predicted feasible strategy.

[0040] The determination manner of the predicted charging demand feature of the charging pile cluster: obtaining a charging pile cluster in the previous continuous determination of a plurality of actual charging demand features (the form of each actual charging demand feature is {the number of vehicles to be charged, the proportion of emergency charging demand, the demand power distribution, and the overflow demand of adjacent clusters}), and integrating the plurality of actual charging demand features into an actual charging demand time series feature set in a time series set manner. The actual charging demand time series feature set is input into the cluster charging demand prediction model corresponding to the charging pile cluster. The cluster charging demand prediction model outputs the predicted charging demand feature of the charging pile cluster;

[0041] The format of the actual charging demand time series feature set: JSON array format:

[0042] { / * 2025-06-15T08:00:00 number of vehicles to be charged, proportion of emergency charging demand, demand power distribution, and overflow demand of adjacent clusters * / },

[0043] { / * 2025-06-15T08:15:00 number of vehicles to be charged, proportion of emergency charging demand, demand power distribution, and overflow demand of adjacent clusters * / },

[0044] { / * 2025-06-15T08:30:00 number of vehicles to be charged, proportion of emergency charging demand, demand power distribution, and overflow demand of adjacent clusters * / },

[0045] ...].

[0046] Each charging pile cluster corresponds to a cluster charging demand prediction model, all cluster charging demand prediction models are built based on LSTM model, each cluster charging demand prediction model exists on the difference of training data, the training process is similar, this embodiment takes charging pile cluster A as an example, the building method of cluster charging demand prediction model is disclosed: a plurality of actual charging demand time sequence feature sets of charging pile cluster A are collected, an LSTM model is built (the [time step, feature number] is specified, a Dropout layer is added to prevent overfitting, and the output layer is designed to be consistent with the input feature number), the actual charging demand time sequence feature set is used as basic data, the LSTM model is trained, and each actual charging demand time sequence feature set is assigned a predicted charging demand feature (the assignment process includes normalizing each feature in the actual charging demand time sequence feature set, for example, the original value range of the feature name: the number of vehicles to be charged is 8-12, and the normalized range is 0-1, the original value range of the feature name: the proportion of emergency charging demand is 0.2-0.3, and the normalized range is 0-1, the original value range of the feature name: adjacent overflow demand is 1-3, and the normalized range is 0-1, the normalization process is not expanded, such as using Min-Max standardization method), the predicted charging demand feature is the actual charging demand feature of the next period of charging pile cluster A, then a plurality of actual charging demand time sequence feature sets are divided into training set, validation set and test set according to a certain proportion, the specific division proportion is determined as 70:15:15, the training set is used to repeatedly train the LSTM model, and the model training key parameters are as follows:

[0047] model.compile(

[0048] optimizer='adam',

[0049] loss='mse', # Mean Squared Error loss function is suitable for regression task

[0050] metrics=['mae'] # Mean Absolute Error for intuitive evaluation)

[0051] # Training configuration

[0052] history = model.fit(

[0053] X_train, y_train,

[0054] epochs=100,

[0055] batch_size=32,

[0056] validation_split=0.15,

[0057] callbacks=[

[0058] EarlyStopping(patience=10, restore_best_weights=True),

[0059] ReduceLROnPlateau(factor=0.5, patience=5)

[0060] The model uses a validation set to verify its performance during the training phase. Based on the validation results, the model parameters are adjusted in a timely manner. Hyperparameter tuning, overfitting prevention, and training monitoring are employed. The final model is evaluated using a test set that was not used in the training process to ensure that the results do not depend on data snooping during the training process. Finally, the cluster charging demand prediction model for charging pile cluster A is completed.

[0061] The charging allocation and adjustment determination unit matches each currently feasible strategy with each predicted feasible strategy, marks the matched currently feasible strategies and predicted feasible strategies as a strategy connection combination, further determines the dynamic connection index of each strategy connection combination, further determines the reasonable value of the preceding connection for each currently feasible strategy, marks the currently feasible strategy with the largest reasonable value of the preceding connection as the currently determined strategy, and uses the currently determined strategy to allocate and adjust the charging of all electric two-wheeled vehicles in the charging pile cluster within the charging area.

[0062] The method for determining the dynamic integration index of strategy integration combination: determine the feature set of currently feasible strategies in the strategy integration combination. (For example, a1 corresponds to the total power limit of charging pile cluster A of 120kW, a2 corresponds to the total power limit of charging pile cluster B of 80kW, a3 corresponds to the fast charging power of charging pile cluster A of 80kW, and a4 corresponds to the slow charging power of charging pile cluster A of 40kW), determine the feature set of predicted feasible strategies in the strategy connection combination. (For example, b1 corresponds to the total power limit of charging pile cluster A of 80W, b2 corresponds to the total power limit of charging pile cluster B of 160kW, b3 corresponds to the fast charging power of charging pile cluster A of 32kW, and b4 corresponds to the slow charging power of charging pile cluster A of 48kW), i represents the i-th feature in the current feasible strategy or the predicted feasible strategy. The similarity between the two feature sets is calculated using the cosine similarity formula. The calculated similarity is the dynamic connection index of the strategy connection combination (e.g., only a1 to a4 and b1 to b4 are calculated). * =120×80+80×160+80×32+40×48=9600+12800+2560+1920=26880;

[0063] || || ≈169.7;

[0064] || || ≈187.96, the cosine similarity calculates the dynamic connection index of the strategy connection combination as 0.84).

[0065] The determination method of the reasonable value of the front connection of the current feasible strategy is as follows: a current feasible strategy is selected, all strategy connection combinations containing the current feasible strategy are marked as possible connection combinations, the dynamic connection indexes of all possible connection combinations are calculated, the average dynamic connection index is calculated, all possible connection combinations are matched two by two, the dynamic connection difference indexes of the two matched possible connection combinations are calculated, the average dynamic connection difference index is calculated, and the average dynamic connection index is divided by the average dynamic connection difference index to calculate the reasonable value of the front connection of the current feasible strategy.

[0066] The above system realizes digital modeling and dynamic correlation analysis of the electric two-wheeled vehicle charging pile cluster by building a virtual cluster system, relies on multi-dimensional data integration of geographic distribution, capacity scale and electricity consumption attributes, combines spatial correlation and load correlation comprehensive computing mechanism, forms a digital twin model that can map the physical cluster state in real time, learns the historical charging demand time series characteristics through the LSTM model, and combines the quantitative analysis of dynamic connection indexes and front connection reasonable values to build a smooth transition mechanism for the current strategy and the predicted strategy, ensuring that the current charging allocation and adjustment strategy not only meets the current demand, but also maximizes the convenience of the connection and switching of subsequent strategies.

[0067] Embodiment two: an electric two-wheeled vehicle charging method based on dynamic allocation and adjustment, the steps are as follows:

[0068] Step one: build a virtual cluster system;

[0069] Step two: periodically determine the current feasible strategy and the predicted feasible strategy by the virtual cluster system;

[0070] Step three: select a current determination strategy from the current feasible strategy, and use the current determination strategy to allocate and adjust the charging of all electric two-wheeled vehicle charging pile clusters in the charging area.

[0071] The above method provides core technical support for the intelligent upgrading of urban micro-mobility charging networks through the closed-loop control of "real-time demand analysis-future trend prediction-strategy connection evaluation".

[0072] The above formulas are all de-dimensioned to calculate the numerical values, the formulas are obtained by collecting a large amount of data to simulate a formula of the most recent real situation, and the preset parameters in the formulas are set by a person skilled in the art according to the actual situation.

[0073] The above embodiments can be implemented wholly or partially by software, hardware, firmware, or any other combination. When implemented by software, the above embodiments can be implemented wholly or partially in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another, for example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center through a wired or wireless (such as infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (such as a floppy disk, a hard disk, a magnetic tape), an optical medium (such as a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0074] It should be understood that the size of the sequence number of each process described above in various embodiments of the present application does not mean the order of execution, and the execution order of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0075] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be realized in electronic hardware, or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0076] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0077] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other manners. For example, the division of the above-described device embodiments is only a logical function division, and there can be another division manner for actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between different units, or the among different units, can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0078] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application can be embodied in the form of a software product, and the computer software product is stored in a storage medium, and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, and various other media that can store program codes.

[0079] The above describes only the specific embodiments of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A charging system for electric two-wheeled vehicles based on dynamic allocation and regulation, characterized by, The virtual cluster establishment unit is configured to determine all charging pile clusters contained in the charging area and build a virtual cluster system. The current feasible strategy analysis unit is configured to determine a current feasible strategy based on the periodic virtual cluster system. The predicted feasible strategy analysis unit is configured to determine predicted charging demand characteristics of each charging pile cluster and further determine a predicted feasible strategy. The charging distribution adjustment determination unit is configured to generate a plurality of strategy connection combinations based on the current feasible strategy and the predicted feasible strategy, determine a dynamic connection index of each strategy connection combination, further determine a pre-connection reasonable value of each current feasible strategy, mark a current feasible strategy with the largest pre-connection reasonable value as a current determination strategy, and use the current determination strategy to perform charging distribution and adjustment on all charging pile clusters of electric two-wheeled vehicles in the charging area. The charging distribution adjustment determination unit is configured to generate a plurality of strategy connection combinations based on the current feasible strategy and the predicted feasible strategy, and mark the matched current feasible strategy and predicted feasible strategy as one strategy connection combination. The determination manner of the dynamic connection index of the strategy connection combination: determining the feature set of the current feasible strategy in the strategy connection combination , determining the feature set of the predicted feasible strategy in the strategy connection combination , represents the i-th feature in the current feasible strategy or the predicted feasible strategy, the similarity of the two feature sets is calculated by using the cosine similarity formula, and the calculated similarity is the dynamic connection index of the strategy connection combination; The pre-connection reasonable value of the current feasible strategy is determined by selecting a current feasible strategy, marking all strategy connection combinations containing the current feasible strategy as possible connection combinations, performing sum average calculation on dynamic connection indexes of all possible connection combinations to calculate an average dynamic connection index, performing two-by-two matching on all possible connection combinations, performing absolute difference calculation on dynamic connection indexes of two matched possible connection combinations to calculate a dynamic connection difference index, performing sum average calculation on all dynamic connection difference indexes to calculate an average dynamic connection difference index, and performing ratio calculation on the average dynamic connection index and the average dynamic connection difference index to calculate the pre-connection reasonable value of the current feasible strategy.

2. The dynamically allocated and regulated electric scooter charging system of claim 1, wherein, The current feasible strategy is determined based on the periodic virtual cluster system, and the current feasible strategy is determined based on the periodic virtual cluster system.

3. The dynamically allocated and regulated electric scooter charging system of claim 1, wherein, The predicted feasible strategy is determined by converting the predicted charging demand characteristics of each charging pile cluster into input parameters and importing them into the virtual cluster system, and the virtual cluster system generates a plurality of charging distribution adjustment strategies, each of which is marked as a predicted feasible strategy.

4. The dynamically allocated and regulated electric scooter charging system of claim 3, wherein, The predicted charging demand characteristics of the charging pile cluster are determined by obtaining a charging pile cluster in a plurality of consecutive actual charging demand characteristics, integrating the plurality of actual charging demand characteristics into an actual charging demand time series feature set in a time series set manner, importing the actual charging demand time series feature set into a cluster charging demand prediction model corresponding to the charging pile cluster, and the cluster charging demand prediction model exports the predicted charging demand characteristics of the charging pile cluster.

5. The method for charging electric two-wheeled vehicles based on dynamic allocation and adjustment, applied to the system for charging electric two-wheeled vehicles based on dynamic allocation and adjustment according to any one of claims 1-4, characterized in that, The steps are as follows: Step one: build a virtual cluster system; Step two: determine a current feasible strategy and a predicted feasible strategy based on a periodic virtual cluster system; Step three: select a present determined strategy from the present feasible strategies, and use the present determined strategy to allocate and adjust the charging of the cluster of charging piles for all electric two-wheel vehicles in the charging area.

6. Electric two-wheeler charging storage medium based on dynamic allocation and regulation, characterized in that, The application is applied to store the electric two-wheel vehicle charging system based on dynamic allocation and adjustment as claimed in any one of claims 1-4.

Images