A rolling optimization method and system for dynamic scheduling of electric vehicle charging and discharging resources

Abstract

The application discloses a rolling optimization electric vehicle charging and discharging resource dynamic scheduling method and system, relates to the technical field of resource scheduling, and comprises the following steps: performing adaptive optimization on a rolling time window according to predicted power grid operation state data and predicted electric vehicle access state data, and obtaining an optimal rolling time window; dividing future time periods according to the optimal rolling time window, and generating an adaptive rolling time window sequence; performing scheduling strategy optimization on charging and discharging resources, and outputting an optimal scheduling strategy for iterative control. The method solves the problem that the existing electric vehicle charging and discharging resource scheduling method is difficult to adapt to the dynamic changes of the power grid operation state and the electric vehicle access state, and the low charging efficiency of the vehicle and the large power grid load fluctuation.

Description

Technical Field

[0001] This application relates to the field of resource scheduling technology, specifically to a rolling optimization method and system for dynamic scheduling of electric vehicle charging and discharging resources. Background Technology

[0002] With the increasing popularity of electric vehicles, their charging and discharging behavior has a significant impact on the power grid. However, traditional electric vehicle charging and discharging resource scheduling methods are difficult to adapt to the dynamic changes in power grid operation and electric vehicle access status, resulting in low vehicle charging efficiency and large fluctuations in power grid load. Summary of the Invention

[0003] This application provides a rolling optimization method and system for dynamic scheduling of electric vehicle charging and discharging resources, which solves the technical problem that existing electric vehicle charging and discharging resource scheduling methods are unable to adapt to the dynamic changes in grid operation status and electric vehicle access status, resulting in low vehicle charging efficiency and large grid load fluctuations.

[0004] The technical solution to the above-mentioned technical problems in this application is as follows:

[0005] Firstly, this application provides a rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources, the method comprising:

[0006] Adaptive optimization of the rolling time window is performed based on the predicted power grid operation status data and predicted electric vehicle access status data for the target area in the future time period to obtain the optimal rolling time window;

[0007] The future time period is divided according to the optimal rolling time window to generate an adapted rolling time window sequence;

[0008] Based on the adapted rolling time window sequence, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, the scheduling strategy for charging and discharging resources is optimized according to the preset monitoring feedback adjustment mechanism, and the optimal scheduling strategy is output for iterative regulation.

[0009] Secondly, this application provides a rolling optimization dynamic scheduling system for electric vehicle charging and discharging resources, comprising:

[0010] The adaptive optimization module is used to adaptively optimize the rolling time window based on the predicted power grid operation status data and predicted electric vehicle access status data of the target area in the future time period, so as to obtain the optimal rolling time window.

[0011] The sequence generation module is used to divide the future time period according to the optimal rolling time window and generate a sequence adapted to the rolling time window.

[0012] The strategy optimization module is used to optimize the scheduling strategy of charging and discharging resources based on the adapted rolling time window sequence, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, according to a preset monitoring feedback adjustment mechanism, and output the optimal scheduling strategy for iterative regulation.

[0013] This application provides one or more technical solutions, which have at least the following technical effects or advantages:

[0014] This application provides a rolling optimization method and system for dynamic scheduling of electric vehicle charging and discharging resources. First, adaptive optimization of the rolling time window is performed to obtain the optimal rolling time window, enabling scheduling to better align with the actual dynamic changes in the power grid and electric vehicle access. Second, the optimal rolling time window is used to divide future time periods, generating an adaptive rolling time window sequence, providing a reasonable time framework for subsequent scheduling strategy optimization. Simultaneously, during the optimization of the charging and discharging resource scheduling strategy, the goal is to maximize vehicle charging efficiency and minimize power grid load fluctuations. A preset monitoring and feedback adjustment mechanism is implemented, continuously monitoring and providing feedback to adjust the scheduling strategy in a timely manner, ensuring efficient charging and discharging resource scheduling under different power grid operating conditions and electric vehicle access scenarios. Finally, iterative adjustments are made based on real-time changes, avoiding the limitations of traditional methods in the face of dynamic changes, thereby effectively improving vehicle charging efficiency and reducing power grid load fluctuations.

[0015] Through the above technical solution, this application combines rolling optimization with distributed computing to realize the continuous evolution and closed-loop control of scheduling strategies in a dynamic environment, effectively improving the response speed and adaptability of electric vehicles participating in grid scheduling. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating a rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the structure of a rolling optimization dynamic scheduling system for electric vehicle charging and discharging resources provided in an embodiment of this application.

[0019] The components represented by each number in the attached diagram are explained below:

[0020] Adaptive optimization module 11, sequence generation module 12, strategy optimization module 13. Detailed Implementation

[0021] This application provides a rolling optimization method and system for dynamic scheduling of electric vehicle charging and discharging resources, which addresses the technical problem that existing electric vehicle charging and discharging resource scheduling methods are unable to adapt to the dynamic changes in grid operation status and electric vehicle access status, resulting in low vehicle charging efficiency and large grid load fluctuations.

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0024] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0025] Example 1, as Figure 1 As shown in the figure, this application provides a rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources, including:

[0026] S10: Adaptively optimize the rolling time window based on the predicted power grid operation status data and predicted electric vehicle access status data for the target area in the future time period to obtain the optimal rolling time window;

[0027] In this embodiment of the application, firstly, the predicted power grid operation status data and the predicted electric vehicle access status data of the target area in the future time period are collected. The predicted power grid operation status data includes indicators such as power grid voltage, frequency, and power, and the predicted electric vehicle access status data includes access time, number of accesses, charging demand, etc.

[0028] Secondly, adaptive optimization of the rolling time window is performed based on the collected data, taking into account peak and off-peak periods of the power grid and the concentrated access time of electric vehicles. For example, during peak power consumption periods, the rolling time window can be appropriately shortened to avoid excessive grid load; while during off-peak periods, the rolling time window can be appropriately extended to improve the flexibility of dispatching.

[0029] Specifically, to achieve adaptive optimization of the rolling time window, data analysis and machine learning algorithms are used. Historical power grid operation status data and electric vehicle access status data are used as training samples to train a prediction model, providing a basis for the optimization of the rolling time window.

[0030] At the same time, a feedback mechanism is introduced to dynamically adjust the rolling time window based on the actual scheduling effect, ensuring that it always adapts to the actual operation of the power grid and electric vehicles.

[0031] Obtaining the optimal rolling time window provides a foundation for subsequent electric vehicle charging and discharging resource scheduling. Scheduling within the optimal time window allows for a better balance between vehicle charging efficiency and grid load fluctuations, achieving optimal resource allocation.

[0032] This includes acquiring predicted power grid operation status data and predicted electric vehicle access status data for the target area in the future time period, including:

[0033] Obtain historical power grid load sequences and historical vehicle charging load sequences for the target area within a historical time period;

[0034] Based on a long short-term memory network, and according to the historical power grid load sequence and the historical vehicle charging load sequence, a predicted power grid load sequence and a predicted vehicle charging load sequence for a future period are obtained as predicted power grid operating status data and predicted electric vehicle access status data, wherein the time interval of the future period is the same as the time interval of the historical period.

[0035] In this embodiment, firstly, historical power grid load sequences and historical vehicle charging load sequences for the target area within a historical time period are collected. These can be obtained from sources such as power grid monitoring systems and electric vehicle charging management systems. Long Short-Term Memory (LSTM) networks are specialized recurrent neural networks capable of handling long-term dependencies in sequence data, making them suitable for load forecasting.

[0036] Secondly, the collected historical power grid load sequences and historical vehicle charging load sequences are input into a trained LSTM model. This model predicts the power grid load and vehicle charging load in future time periods based on the characteristics and patterns of the data, thus obtaining predicted power grid load sequences and predicted vehicle charging load sequences. Since the time intervals between future and historical periods are the same, the accuracy and comparability of the predictions are ensured, allowing subsequent scheduling strategies to better adapt to actual conditions.

[0037] Furthermore, after obtaining the predicted grid load sequence and the predicted vehicle charging load sequence, adaptive optimization of the rolling time window is performed, and then subsequent optimization of charging and discharging resource scheduling strategies is carried out to achieve the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations.

[0038] Furthermore, based on a Long Short-Term Memory (LSTM) network, and according to the historical power grid load sequence and historical vehicle charging load sequence, a predicted power grid load sequence and a predicted vehicle charging load sequence for future periods are obtained, including:

[0039] Based on the historical power grid operation records of the target area, and constrained by the historical time period, a sample power grid load sequence set is collected, and the historical power grid load sequences of different sample power grid load sequences in the historical future time period are collected as sample predicted power grid load sequences to obtain a sample predicted power grid load sequence set.

[0040] Using the sample power grid load sequence set as input and the sample predicted power grid load sequence set as supervision, a long short-term memory network is trained until convergence to generate a power grid load predictor.

[0041] The power grid load forecaster is used to predict the power grid load sequence for future periods based on the historical power grid load sequence.

[0042] Based on the historical power grid operation records of the target area, and constrained by the historical time period, a sample vehicle charging load sequence set and a sample predicted vehicle charging load sequence set are collected.

[0043] Using the sample vehicle charging load sequence set as input and the sample predicted vehicle charging load sequence set as supervision, a long short-term memory network is trained until convergence to generate a vehicle charging load predictor.

[0044] The vehicle charging load predictor uses the historical vehicle charging load sequence to predict the vehicle charging load sequence for future periods.

[0045] In this embodiment, firstly, based on the historical power grid operation records of the target area, a historical time period is determined, and multiple sample power grid load sequences are collected within this range to form a sample power grid load sequence set. Simultaneously, for each sample power grid load sequence, its historical power grid load sequence within the corresponding historical future time period is collected to form a sample predicted power grid load sequence set.

[0046] Secondly, the sample power grid load sequence set is used as input data, and the sample predicted power grid load sequence set is used as supervision data to train the Long Short-Term Memory (LSTM) network. During training, the network continuously adjusts its weights and parameters to make the predicted output as close as possible to the supervision data. When the network's loss function converges to a small value, the training is considered complete, and a power grid load predictor capable of accurately predicting power grid load is generated.

[0047] Then, the historical grid load sequence of the target area is input into the trained grid load predictor. Based on the learned patterns, the predictor outputs the predicted grid load sequence for future periods. The predicted sequence reflects the general trend of future grid load, providing a basis for subsequent rolling time window adaptive optimization and charging / discharging resource scheduling.

[0048] For example, a power grid load forecaster is built and trained based on a long short-term memory network. The specific steps are as follows:

[0049] First, data preparation involves collecting sample power grid load sequences and historical power grid load sequences of different sample power grid load sequences within historical future time periods, based on historical power grid operation records of the target area.

[0050] Secondly, the model is built using a sample set of power grid load sequences as input and the predicted power grid load as output. The Long Short-Term Memory (LSTM) network is implemented using the TensorFlow framework. The number of nodes in the input layer equals the dimension of the input features. For example, if the sample predicted power grid load sequence has two features, the input layer contains two nodes. One to three hidden layers are set, with the number of nodes in each layer adjusted experimentally (e.g., 64, 32, etc.). The ReLU activation function is used. The number of nodes in the output layer equals the number of predicted power grid loads. For example, if the prediction time requires one node, the output layer generally does not use an activation function and directly outputs continuous values.

[0051] Then, the model is trained. In each training iteration, the sample predicted grid load sequence set is used as supervision. The Adam optimizer and mean square error (MSE) loss function are used to build the training framework. The batch size is set to 32 and the total number of training rounds is 50. An early stop mechanism is introduced with patience=5. When the validation set loss does not decrease for 5 consecutive rounds, the training process is automatically terminated, and the trained grid load predictor is obtained.

[0052] Furthermore, for predicting vehicle charging load sequences, based on historical power grid operation records of the target area, multiple sample vehicle charging load sequences are collected under the constraint of historical time periods to form a sample vehicle charging load sequence set. Simultaneously, historical vehicle charging load sequences of the samples within historical future time periods are collected to form a sample predicted vehicle charging load sequence set.

[0053] Next, the same method used to train the power grid load predictor is employed to train the vehicle charging load predictor. The sample vehicle charging load sequence set is used as input, and the sample predicted vehicle charging load sequence set is used as supervision to train the Long Short-Term Memory (LSTM) network. The network parameters are continuously adjusted until the network converges, thus generating the vehicle charging load predictor.

[0054] Finally, the historical vehicle charging load sequence is input into the vehicle charging load predictor to predict the vehicle charging load sequence for future periods. By combining the predicted grid load sequence and the predicted vehicle charging load sequence, the operating status of the power grid and the access status of electric vehicles in the target area in the future period can be understood, thereby enabling adaptive optimization of the rolling time window and optimization of the scheduling strategy for charging and discharging resources.

[0055] Specifically, adaptive optimization of the scrolling time window is performed to obtain the optimal scrolling time window, including:

[0056] The load volatility is calculated based on the predicted power grid load sequence to obtain the power grid load volatility, wherein the power grid load volatility is the ratio of the standard deviation to the mean of the predicted power grid load in the predicted power grid load sequence;

[0057] The load variability is calculated based on the predicted vehicle charging load sequence to obtain the charging load variability.

[0058] The vehicle charging load percentage sequence is calculated based on the predicted power grid load sequence and the predicted vehicle charging load sequence, and the average load percentage and load percentage fluctuation are calculated.

[0059] Adaptive optimization of the rolling time window is performed based on the grid load fluctuation, charging load fluctuation, average load ratio, and load ratio fluctuation to obtain the optimal rolling time window.

[0060] In this embodiment, firstly, load volatility is calculated on the predicted power grid load sequence. The ratio of the standard deviation to the mean of the predicted power grid load is calculated to obtain the power grid load volatility. This index reflects the degree of fluctuation of the power grid load in the future period; the greater the volatility, the more unstable the power grid load.

[0061] Secondly, the same load volatility calculation is performed on the predicted vehicle charging load sequence to obtain the charging load volatility, which is used to measure the volatility of electric vehicle charging load.

[0062] Next, the vehicle charging load share sequence is calculated, which predicts the proportion of vehicle charging load in the total grid load. Then, the mean and volatility of this sequence are calculated to obtain the load share mean and load share volatility, respectively. The load share mean reflects the average level of electric vehicle charging load in the total grid load, while the load share volatility reflects the fluctuation of this proportion.

[0063] For example, the predicted grid load sequence is [100, 120, 110, 130, 140], and the predicted vehicle charging load sequence is [20, 25, 22, 28, 30]. The calculated mean of the grid load is (100+120+110+130+140) / 5=120, and the standard deviation is approximately 14.14. Therefore, the grid load fluctuation is approximately 14.14 / 120≈0.12.

[0064] The mean of the charging load is (20+25+22+28+30) / 5=25, the standard deviation is approximately 3.58, and the charging load fluctuation is approximately 3.58 / 25=0.14. The vehicle charging load percentage sequences are 20 / 100=0.2, 25 / 120≈0.21, 22 / 110=0.2, 28 / 130≈0.22, 30 / 140≈0.21, and the mean load percentage is approximately (0.2+0.21+0.2+0.22+0.21) / 5=0.206. The load percentage fluctuation can be obtained by calculating the ratio of the standard deviation to the mean of this sequence.

[0065] Furthermore, based on the above-calculated grid load fluctuation, charging load fluctuation, average load percentage, and load percentage fluctuation, adaptive optimization of the rolling time window is performed.

[0066] For example, when the grid load fluctuates significantly, the rolling time window is narrowed to respond promptly to changes in the grid load; when the charging load fluctuates significantly, the rolling time window is adjusted according to the actual situation to ensure that the charging needs of electric vehicles can be reasonably met.

[0067] Meanwhile, by combining the average load share and the load share fluctuation, and considering the position and changes of electric vehicle charging load in the power grid, the optimal rolling time window is determined.

[0068] Through the above optimization methods, the rolling time window can be dynamically adjusted according to the actual operating status of the power grid and electric vehicles, providing a more suitable time frame for subsequent charging and discharging resource scheduling, thereby improving the efficiency and adaptability of scheduling and better achieving the goal of maximizing vehicle charging efficiency and minimizing power grid load fluctuations.

[0069] Furthermore, based on the grid load fluctuation, charging load fluctuation, average load percentage, and load percentage fluctuation, adaptive optimization of the rolling time window is performed to obtain the optimal rolling time window, including:

[0070] The ratio of the preset standard power grid load fluctuation to the power grid load fluctuation is set as the first window compensation coefficient;

[0071] The ratio of the preset standard charging load fluctuation to the charging load fluctuation is set as the second window compensation coefficient.

[0072] The ratio of the preset standard load percentage average to the load percentage average is set as the third window compensation coefficient;

[0073] The ratio of the preset standard load percentage fluctuation to the load percentage fluctuation is set as the fourth window compensation coefficient.

[0074] The first window compensation coefficient, the second window compensation coefficient, the third window compensation coefficient, and the fourth window compensation coefficient are weighted and fused to generate a comprehensive window compensation coefficient;

[0075] Based on the optimization and correction of the initial scrolling time window using the comprehensive window compensation coefficient, the optimal scrolling time window is output.

[0076] In this embodiment, firstly, preset standard grid load fluctuation, preset standard charging load fluctuation, preset standard load percentage average, and preset standard load percentage fluctuation are determined. These preset standard values ​​are determined based on historical data statistical analysis, grid planning requirements, and electric vehicle development trends, representing the fluctuation and percentage of grid and electric vehicle charging loads under ideal conditions.

[0077] Next, the compensation coefficients for the first window, second window, third window, and fourth window are calculated respectively. The compensation coefficient for the first window reflects the degree of difference between the actual power grid load fluctuation and the preset standard. If the coefficient is greater than 1, it indicates that the actual power grid load fluctuation is less than the preset standard, and the power grid operation is relatively stable; if it is less than 1, it indicates that the actual fluctuation is large and the power grid stability is poor.

[0078] Similarly, the second window compensation coefficient reflects the comparison between the electric vehicle charging load fluctuation and the standard, the third window compensation coefficient reflects the deviation of the average electric vehicle charging load ratio, and the fourth window compensation coefficient measures the difference in load ratio fluctuation.

[0079] Then, the four window compensation coefficients are weighted and integrated. The weight allocation is determined based on the actual needs and priorities of the power grid operation. For example, if the power grid is more sensitive to load fluctuations, then the first and second window compensation coefficients are assigned larger weights; if more attention is paid to the proportion of electric vehicle charging load in the power grid, then the weights of the third and fourth window compensation coefficients are appropriately increased. Through weighted integration, the four coefficients are combined into a comprehensive window compensation coefficient, which can comprehensively reflect the deviation of the actual operating status of the power grid and electric vehicle charging load from the preset standard.

[0080] For example, assume that the first window compensation coefficient is 0.9, the second window compensation coefficient is 1.1, the third window compensation coefficient is 0.8, and the fourth window compensation coefficient is 1.2. Based on the actual needs of power grid operation, it is more sensitive to load fluctuations. The weights assigned to the first window compensation coefficient and the second window compensation coefficient are both 0.3, and the weights assigned to the third window compensation coefficient and the fourth window compensation coefficient are both 0.2.

[0081] The comprehensive window compensation coefficient is 0.9×0.3+1.1×0.3+0.8×0.2+1.2×0.2=1.0.

[0082] Finally, the initial rolling time window is optimized and corrected based on the comprehensive window compensation coefficient. If the comprehensive window compensation coefficient is greater than 1, it indicates that the actual operating condition is better than the preset standard, and the initial rolling time window can be appropriately expanded to improve the scheduling flexibility and resource utilization efficiency. If the comprehensive window compensation coefficient is less than 1, it means that the actual situation is relatively poor, and the initial rolling time window needs to be narrowed to respond more promptly to changes in the grid and electric vehicle charging load.

[0083] After optimization and correction, the output optimal rolling time window can better adapt to the actual operating conditions of the power grid and electric vehicles, providing a precise time frame for subsequent electric vehicle charging and discharging resource scheduling, thereby effectively achieving the goal of maximizing vehicle charging efficiency and minimizing power grid load fluctuations.

[0084] S20: Divide the future time period according to the optimal rolling time window and generate an adapted rolling time window sequence;

[0085] In this embodiment, firstly, the future time period is evenly divided into multiple sub-time periods according to the size of the optimal rolling time window. The length of each sub-time period is the length of the optimal rolling time window, and the sub-time periods are arranged sequentially to form an adaptive rolling time window sequence. Each rolling time window in this sequence corresponds to a specific time period within the future time period, providing a clear time range for subsequent electric vehicle charging and discharging resource scheduling.

[0086] For example, if the optimal rolling time window is 1 hour and the future period is the next 24 hours, then the 24 hours are divided into 24 1-hour sub-periods, forming an adaptive rolling time window sequence containing 24 rolling time windows. Within each rolling time window, scheduling decisions for electric vehicle charging and discharging resources can be made independently to adapt to the dynamic changes in grid load and electric vehicle charging demand.

[0087] Furthermore, new rolling time windows continuously enter the scheduling scope, while old rolling time windows exit. This rolling scheduling method can take into account the latest grid load forecasts and electric vehicle charging load forecasts in real time, thereby adjusting scheduling strategies promptly and improving the accuracy and effectiveness of scheduling.

[0088] S30: Based on the adapted rolling time window sequence, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, the scheduling strategy for charging and discharging resources is optimized according to the preset monitoring feedback adjustment mechanism, and the optimal scheduling strategy is output for iterative regulation.

[0089] In this embodiment, firstly, a preset monitoring feedback adjustment mechanism monitors the actual situation of the power grid load and electric vehicle charging load in real time within each adaptive rolling time window. The monitoring content includes the actual power grid load value, changes in electric vehicle charging demand, charging power, etc.

[0090] Secondly, the actual data obtained from monitoring is compared and analyzed with the previous forecast data. If there is a deviation between the actual grid load and the forecast grid load, or a deviation between the actual charging demand of electric vehicles and the forecast charging load, the scheduling strategy needs to be adjusted according to the deviation.

[0091] Within each rolling time window, an optimization model is constructed with the objectives of maximizing vehicle charging efficiency and minimizing grid load fluctuations. This model considers factors such as the battery state of the electric vehicle, charging priority, and real-time electricity prices. Through analysis and calculation, the optimal scheduling strategy within the current rolling time window is sought.

[0092] For example, when the grid load is low and the electricity price is cheap, more electric vehicles are prioritized for charging to improve charging efficiency; while when the grid load is high, the charging power is appropriately reduced or the charging time is adjusted to reduce grid load fluctuations.

[0093] During the optimization process, an iterative approach is used to continuously optimize the scheduling strategy. Each iteration adjusts and improves the optimization model based on the previous scheduling results and monitoring feedback. Through multiple iterations, the scheduling strategy gradually approaches the optimal solution.

[0094] Meanwhile, the dynamic access and exit of electric vehicles should be taken into account. When a new electric vehicle is connected to the grid, it needs to be included in the dispatch scope in a timely manner, and the dispatch strategy should be adjusted according to its charging demand and battery status; when an electric vehicle completes charging and leaves, the charging arrangements for the remaining electric vehicles should be adjusted accordingly.

[0095] Through continuous optimization and iterative adjustments, the final optimal scheduling strategy can meet the charging needs of electric vehicles while minimizing grid load fluctuations, thus achieving coordinated and optimized operation of the grid and electric vehicles.

[0096] Specifically, step S30 in the method includes:

[0097] In the sequence of adaptive scrolling time windows, the first adaptive scrolling time window is selected as the first adaptive scrolling time window, and the adjacent window of the first adaptive scrolling time window is selected as the second adaptive scrolling time window.

[0098] The first historical power grid load sequence and the first historical vehicle charging load sequence within the first adaptation rolling time window are monitored and obtained.

[0099] A short-term power grid load predictor and a short-term vehicle charging load predictor are constructed based on a long short-term memory network. A second predicted power grid load sequence and a second predicted vehicle charging load sequence are obtained based on the first historical power grid load sequence and the first historical vehicle charging load sequence.

[0100] With the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, the scheduling strategy for charging and discharging resources is optimized according to a preset monitoring feedback adjustment mechanism, and the second optimal scheduling strategy is output.

[0101] The second optimal scheduling strategy is used to schedule electric vehicle charging and discharging resources within the second adaptive rolling time window, and iterative regulation is performed based on the adaptive rolling time window sequence.

[0102] In this embodiment, firstly, the first window is selected from the adaptive rolling time window sequence as the first adaptive rolling time window, and its adjacent windows are selected as the second adaptive rolling time window. Based on the actual situation of the current window, the scheduling of adjacent windows is predicted and adjusted to achieve rolling dynamic scheduling.

[0103] Secondly, the actual operating data within the first adaptive rolling time window is monitored to obtain the first historical power grid load sequence and the first historical electric vehicle charging load sequence. The first historical power grid load sequence and the first historical electric vehicle charging load sequence contain the actual changes in the power grid and electric vehicle charging load within this time window.

[0104] Then, a short-term grid load forecaster and a short-term vehicle charging load forecaster are constructed using a Long Short-Term Memory (LSTM) network. Compared with the previously mentioned grid load forecaster and vehicle charging load forecaster, the grid load forecaster and the short-term vehicle charging load forecaster have shorter time periods, thus achieving higher prediction accuracy. Based on the first historical grid load sequence and the first historical vehicle charging load sequence, a second predicted grid load sequence and a second predicted vehicle charging load sequence are obtained through the grid load forecaster and the short-term vehicle charging load forecaster.

[0105] Subsequently, with the goals of maximizing vehicle charging efficiency and minimizing grid load fluctuations, an optimal scheduling strategy for charging and discharging resources is developed by combining a second predicted grid load sequence and a second predicted vehicle charging load sequence, according to a preset monitoring and feedback adjustment mechanism. During the optimization process, factors such as the battery status of electric vehicles, charging priority, and real-time grid electricity prices are considered.

[0106] For example, when it is predicted that the grid load will increase and the electricity price will rise, the charging power of electric vehicles will be reduced or the charging time will be adjusted; when it is predicted that the grid load will be low and the electricity price will be cheap, the charging power will be increased or more electric vehicles will be scheduled to charge. Through continuous analysis and calculation, the optimal scheduling strategy under the current situation is found, and a second optimal scheduling strategy is output.

[0107] Finally, the electric vehicle charging and discharging resource scheduling for the second adaptive rolling time window is executed according to the second optimal scheduling strategy. Furthermore, iterative adjustments are made based on the entire adaptive rolling time window sequence.

[0108] In each subsequent rolling time window, the above steps are repeated, and the scheduling strategy is continuously adjusted based on actual monitoring data and forecast results. This allows the scheduling strategy to adapt to the dynamic changes of the power grid and electric vehicles in real time, ultimately achieving coordinated and optimized operation of the power grid and electric vehicles. This meets the charging needs of electric vehicles while minimizing fluctuations in the power grid load.

[0109] Specifically, with the objectives of maximizing vehicle charging efficiency and minimizing grid load fluctuations, based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, the scheduling strategy for charging and discharging resources is optimized according to a preset monitoring feedback adjustment mechanism, and a second optimal scheduling strategy is output, including:

[0110] Obtain the vehicle charging power adjustment space of the second adaptive rolling time window, and randomly select several initial charging powers within the vehicle charging power adjustment space.

[0111] Based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, grid load is predicted according to the plurality of initial charging powers, and a plurality of predicted grid load fluctuations are output.

[0112] Based on the second predicted vehicle charging load sequence, vehicle charging simulation is performed according to the several initial charging powers, and several simulated vehicle charging efficiencies are output.

[0113] With the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, several parameter fitnesss are calculated by weighting several predicted grid load fluctuations and several simulated vehicle charging efficiencies. The parameter fitnesss are negatively correlated with the predicted grid load fluctuations and positively correlated with the simulated vehicle charging efficiencies.

[0114] Based on the aforementioned initial charging power and the aforementioned parameter fitness, the scheduling strategy for charging and discharging resources is optimized, and a second optimal scheduling strategy is output.

[0115] In this embodiment, firstly, the vehicle charging power adjustment space within the second adaptive rolling time window is determined. The vehicle charging power adjustment space is limited by factors such as the electric vehicle battery capacity, the upper limit of charging infrastructure power, and the requirements for safe operation of the power grid. Several initial charging powers are randomly selected within the vehicle charging power adjustment space as the starting point for the subsequent optimization process.

[0116] Secondly, based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, grid load forecasting is performed for each initial charging power. By constructing a grid load forecasting model, using the initial charging power as input, and combining the predicted grid load and vehicle charging load conditions, several predicted grid load fluctuations are calculated. The predicted grid load fluctuations reflect the magnitude of grid load changes under different initial charging powers and are an indicator of grid stability.

[0117] Simultaneously, based on the second predicted vehicle charging load sequence, vehicle charging simulations are performed for each initial charging power. The simulation considers factors such as the electric vehicle's battery characteristics and charging efficiency curves, outputting several simulated vehicle charging efficiencies. These simulated charging efficiencies reflect the charging performance of the electric vehicle under different initial charging powers.

[0118] Then, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, a weighted calculation is performed on several predicted grid load fluctuations and several simulated vehicle charging efficiencies. The weights are determined based on the actual needs and priorities of grid operation.

[0119] For example, if the power grid prioritizes stability, then a larger weight is given to the predicted power grid load fluctuation; if the focus is more on the charging efficiency of electric vehicles, then the weight of the simulated vehicle charging efficiency is appropriately increased. Several parameter fitness values ​​are obtained through weighted calculation. The parameter fitness values ​​are negatively correlated with the predicted power grid load fluctuation and positively correlated with the simulated vehicle charging efficiency, reflecting the performance of different initial charging powers in terms of target optimization.

[0120] For example, assume that the weight assigned to the predicted grid load fluctuation is 0.6 and the weight assigned to the simulated car charging efficiency is 0.4. There are three initial charging powers, with corresponding predicted grid load fluctuations of 0.2, 0.3, and 0.1, and simulated car charging efficiencies of 0.8, 0.7, and 0.9, respectively.

[0121] The fitness of the parameters corresponding to the first initial charging power is (1-0.2)×0.6+0.8×0.4=0.8;

[0122] The fitness of the parameters corresponding to the second initial charging power is (1-0.3)×0.6+0.7×0.4=0.7;

[0123] The fitness of the parameters corresponding to the third initial charging power is (1-0.1)×0.6+0.9×0.4=0.9.

[0124] Finally, an optimal scheduling strategy for charging and discharging resources is sought based on several initial charging powers and several parameter fitness values. Optimization algorithms such as genetic algorithms and particle swarm optimization can be used to find the charging power combination that maximizes the parameter fitness from the set of initial charging powers, thereby outputting the second optimal scheduling strategy.

[0125] The optimal scheduling strategy can meet the charging needs of electric vehicles while minimizing the fluctuations in grid load, thus achieving coordinated and optimized operation of the grid and electric vehicles.

[0126] Furthermore, based on the aforementioned initial charging power and the aforementioned parameter fitness, a scheduling strategy for charging and discharging resources is optimized, and a second optimal scheduling strategy is output, including:

[0127] Set the initial charging power as the initial solution, and arrange several initial solutions in descending order of parameter fitness to generate an initial solution sequence;

[0128] The first solution in the initial solution sequence is selected as the optimal solution, and the remaining initial solutions are selected as the inferior solutions. The multiple inferior solutions are adjusted according to a preset step size, with the optimal solution as the direction, to obtain multiple updated inferior solutions.

[0129] The optimal solutions and multiple updated inferior solutions are reordered according to the parameter fitness from large to small. After eliminating a preset proportion of the updated inferior solutions, an initial solution is randomly selected in the vehicle charging power adjustment space for equivalent supplementation to construct an updated solution sequence.

[0130] Based on the updated solution sequence, iterative optimization continues until a preset number of convergences is reached. The optimal solution of the current updated solution sequence is then set as the second optimal scheduling strategy.

[0131] In this embodiment of the application, the initial charging power is first used as the initial solution, and the initial solutions are arranged in descending order of parameter fitness to generate an initial solution sequence.

[0132] Secondly, the first solution in the initial solution sequence is identified as the optimal solution, while the remaining initial solutions are considered inferior solutions. Using the optimal solution as a guide, multiple inferior solutions are adjusted according to a preset step size. The preset step size needs to be set reasonably based on the actual situation; too large a step size may lead to over-adjustment and missing the optimal solution; too small a step size will slow down the optimization process. Through adjusting the inferior solutions, multiple updated inferior solutions are obtained.

[0133] Next, the excellent solutions and multiple updated inferior solutions are reordered in descending order of parameter fitness. After sorting, a preset proportion of inferior solutions are eliminated. The preset proportion is determined by comprehensively considering the number of solutions and the efficiency of optimization. After eliminating poorly performing solutions, to ensure the stability of the number of solutions, initial solutions are randomly selected from the vehicle charging power adjustment space for equivalent replenishment, thereby constructing an updated solution sequence.

[0134] Furthermore, iterative optimization continues based on the updated solution sequence. In each iteration, the steps of adjustment, sorting, elimination, and replenishment are repeated. Each iteration optimizes and improves the solution based on the previous result. As the number of iterations increases, the solution gradually approaches the optimal solution.

[0135] When the preset convergence count is reached, it means that the optimization process has basically converged. At this point, the optimal solution of the current updated solution sequence is output and set as the second optimal scheduling strategy. The second optimal scheduling strategy can meet the charging needs of electric vehicles while minimizing the fluctuation of grid load, thus achieving coordinated and optimized operation of the grid and electric vehicles.

[0136] Meanwhile, throughout the optimization process, it is also necessary to pay attention to the dynamic access and exit of electric vehicles, as well as the real-time changes in grid load and electric vehicle charging load, and adjust and optimize the scheduling strategy in a timely manner to ensure the effectiveness and adaptability of the scheduling strategy.

[0137] The preset ratio is the product of the ratio of the remaining number of iterations to the total number of iterations and the initial elimination ratio, wherein the initial elimination ratio is 20%.

[0138] In this embodiment, the preset ratio is the product of the ratio of the remaining iterations to the total number of optimizations and the initial elimination ratio. The number of solutions eliminated is dynamically adjusted during the iterative optimization process. As the remaining iterations decrease, the preset ratio changes accordingly. When there are many remaining iterations, the preset ratio is relatively large, allowing for the rapid elimination of some poorly performing solutions from a large pool of options, thus improving optimization efficiency. Conversely, as iterations progress and the remaining iterations decrease, the preset ratio decreases to avoid excessive elimination of solutions when approaching the optimal solution, which could lead to missing the true optimal solution.

[0139] For example, assuming the total number of optimization attempts is 10, when the third iteration is performed, there are 7 remaining optimization attempts. Then the preset ratio is 7 / 10 × 20% = 14%. At this time, among the reordered excellent solutions and the updated inferior solutions, the last 14% of the updated inferior solutions will be eliminated, and an initial solution will be randomly selected from the car charging power adjustment space to supplement them.

[0140] In summary, compared with existing technologies, this application uses a rolling optimization method to dynamically schedule electric vehicle charging and discharging resources, captures dynamic changes in the power grid and electric vehicles in real time, and adjusts the scheduling strategy accordingly to ensure that the scheduling strategy always matches the actual operating conditions, thereby improving the effectiveness and adaptability of scheduling.

[0141] In summary, the embodiments of this application have at least the following technical effects:

[0142] This application provides a rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources. First, adaptive optimization of the rolling time window is performed to obtain the optimal rolling time window, enabling scheduling to better align with the actual dynamic changes in the power grid and electric vehicle access. Second, the optimal rolling time window is used to divide future time periods, generating an adaptive rolling time window sequence, providing a reasonable time framework for subsequent scheduling strategy optimization. Simultaneously, during the optimization of the charging and discharging resource scheduling strategy, the goal is to maximize vehicle charging efficiency and minimize grid load fluctuations. A preset monitoring and feedback adjustment mechanism is implemented, continuously monitoring and providing feedback to adjust the scheduling strategy in a timely manner, ensuring efficient charging and discharging resource scheduling under different grid operating states and electric vehicle access conditions. Finally, iterative adjustments are made based on real-time changes, avoiding the limitations of traditional methods in the face of dynamic changes, thereby effectively improving vehicle charging efficiency and reducing grid load fluctuations. Through the above technical solution, this application combines rolling optimization with distributed computing to achieve continuous evolution and closed-loop control of the scheduling strategy in a dynamic environment, effectively improving the response speed and adaptability of electric vehicles participating in grid scheduling.

[0143] Example 2, as Figure 2 As shown, based on the same inventive concept as the rolling optimization dynamic scheduling method for electric vehicle charging and discharging resources provided in Embodiment 1, this application also provides a rolling optimization dynamic scheduling system for electric vehicle charging and discharging resources, including:

[0144] The adaptive optimization module 11 is used to adaptively optimize the rolling time window based on the predicted power grid operation status data and the predicted electric vehicle access status data of the target area in the future time period, so as to obtain the optimal rolling time window.

[0145] The sequence generation module 12 is used to divide the future time period according to the optimal rolling time window and generate a sequence adapted to the rolling time window.

[0146] The strategy optimization module 13 is used to optimize the scheduling strategy of charging and discharging resources based on the adapted rolling time window sequence, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, according to a preset monitoring feedback adjustment mechanism, and output the optimal scheduling strategy for iterative regulation.

[0147] Furthermore, in one embodiment of the application, obtaining predicted power grid operating status data and predicted electric vehicle access status data for the target area in a future time period includes:

[0148] Obtain historical power grid load sequences and historical vehicle charging load sequences for the target area within a historical time period;

[0149] Based on a long short-term memory network, and according to the historical power grid load sequence and the historical vehicle charging load sequence, a predicted power grid load sequence and a predicted vehicle charging load sequence for a future period are obtained as predicted power grid operating status data and predicted electric vehicle access status data, wherein the time interval of the future period is the same as the time interval of the historical period.

[0150] Further, in one embodiment of the application, based on a Long Short-Term Memory (LSTM) network, and according to the historical power grid load sequence and the historical vehicle charging load sequence, predicting and obtaining the predicted power grid load sequence and the predicted vehicle charging load sequence for future periods includes:

[0151] Based on the historical power grid operation records of the target area, and constrained by the historical time period, a sample power grid load sequence set is collected, and the historical power grid load sequences of different sample power grid load sequences in the historical future time period are collected as sample predicted power grid load sequences to obtain a sample predicted power grid load sequence set.

[0152] Using the sample power grid load sequence set as input and the sample predicted power grid load sequence set as supervision, a long short-term memory network is trained until convergence to generate a power grid load predictor.

[0153] The power grid load forecaster is used to predict the power grid load sequence for future periods based on the historical power grid load sequence.

[0154] Based on the historical power grid operation records of the target area, and constrained by the historical time period, a sample vehicle charging load sequence set and a sample predicted vehicle charging load sequence set are collected.

[0155] Using the sample vehicle charging load sequence set as input and the sample predicted vehicle charging load sequence set as supervision, a long short-term memory network is trained until convergence to generate a vehicle charging load predictor.

[0156] The vehicle charging load predictor uses the historical vehicle charging load sequence to predict the vehicle charging load sequence for future periods.

[0157] Furthermore, in one embodiment of the application, adaptive optimization of the rolling time window is performed to obtain the optimal rolling time window, including:

[0158] The load volatility is calculated based on the predicted power grid load sequence to obtain the power grid load volatility, wherein the power grid load volatility is the ratio of the standard deviation to the mean of the predicted power grid load in the predicted power grid load sequence;

[0159] The load variability is calculated based on the predicted vehicle charging load sequence to obtain the charging load variability.

[0160] The vehicle charging load percentage sequence is calculated based on the predicted power grid load sequence and the predicted vehicle charging load sequence, and the average load percentage and load percentage fluctuation are calculated.

[0161] Adaptive optimization of the rolling time window is performed based on the grid load fluctuation, charging load fluctuation, average load ratio, and load ratio fluctuation to obtain the optimal rolling time window.

[0162] Furthermore, in one embodiment, adaptive optimization of the rolling time window is performed based on the grid load fluctuation, charging load fluctuation, average load percentage, and load percentage fluctuation to obtain the optimal rolling time window, including:

[0163] The ratio of the preset standard power grid load fluctuation to the power grid load fluctuation is set as the first window compensation coefficient;

[0164] The ratio of the preset standard charging load fluctuation to the charging load fluctuation is set as the second window compensation coefficient.

[0165] The ratio of the preset standard load percentage average to the load percentage average is set as the third window compensation coefficient;

[0166] The ratio of the preset standard load percentage fluctuation to the load percentage fluctuation is set as the fourth window compensation coefficient.

[0167] The first window compensation coefficient, the second window compensation coefficient, the third window compensation coefficient, and the fourth window compensation coefficient are weighted and fused to generate a comprehensive window compensation coefficient;

[0168] Based on the optimization and correction of the initial scrolling time window using the comprehensive window compensation coefficient, the optimal scrolling time window is output.

[0169] In one embodiment, the strategy optimization module 13 is specifically used for:

[0170] In the sequence of adaptive scrolling time windows, the first adaptive scrolling time window is selected as the first adaptive scrolling time window, and the adjacent window of the first adaptive scrolling time window is selected as the second adaptive scrolling time window.

[0171] The first historical power grid load sequence and the first historical vehicle charging load sequence within the first adaptation rolling time window are monitored and obtained.

[0172] A short-term power grid load predictor and a short-term vehicle charging load predictor are constructed based on a long short-term memory network. A second predicted power grid load sequence and a second predicted vehicle charging load sequence are obtained based on the first historical power grid load sequence and the first historical vehicle charging load sequence.

[0173] With the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, the scheduling strategy for charging and discharging resources is optimized according to a preset monitoring feedback adjustment mechanism, and the second optimal scheduling strategy is output.

[0174] The second optimal scheduling strategy is used to schedule electric vehicle charging and discharging resources within the second adaptive rolling time window, and iterative regulation is performed based on the adaptive rolling time window sequence.

[0175] Specifically, with the objectives of maximizing vehicle charging efficiency and minimizing grid load fluctuations, based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, the scheduling strategy for charging and discharging resources is optimized according to a preset monitoring feedback adjustment mechanism, and a second optimal scheduling strategy is output, including:

[0176] Obtain the vehicle charging power adjustment space of the second adaptive rolling time window, and randomly select several initial charging powers within the vehicle charging power adjustment space.

[0177] Based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, grid load is predicted according to the plurality of initial charging powers, and a plurality of predicted grid load fluctuations are output.

[0178] Based on the second predicted vehicle charging load sequence, vehicle charging simulation is performed according to the several initial charging powers, and several simulated vehicle charging efficiencies are output.

[0179] With the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, several parameter fitnesss are calculated by weighting several predicted grid load fluctuations and several simulated vehicle charging efficiencies. The parameter fitnesss are negatively correlated with the predicted grid load fluctuations and positively correlated with the simulated vehicle charging efficiencies.

[0180] Based on the aforementioned initial charging power and the aforementioned parameter fitness, the scheduling strategy for charging and discharging resources is optimized, and a second optimal scheduling strategy is output.

[0181] Furthermore, based on the aforementioned initial charging power and the aforementioned parameter fitness, a scheduling strategy for charging and discharging resources is optimized, and a second optimal scheduling strategy is output, including:

[0182] Set the initial charging power as the initial solution, and arrange several initial solutions in descending order of parameter fitness to generate an initial solution sequence;

[0183] The first solution in the initial solution sequence is selected as the optimal solution, and the remaining initial solutions are selected as the inferior solutions. The multiple inferior solutions are adjusted according to a preset step size, with the optimal solution as the direction, to obtain multiple updated inferior solutions.

[0184] The optimal solutions and multiple updated inferior solutions are reordered according to the parameter fitness from large to small. After eliminating a preset proportion of the updated inferior solutions, an initial solution is randomly selected in the vehicle charging power adjustment space for equivalent supplementation to construct an updated solution sequence.

[0185] Based on the updated solution sequence, iterative optimization continues until a preset number of convergences is reached. The optimal solution of the current updated solution sequence is then set as the second optimal scheduling strategy.

[0186] The preset ratio is the product of the ratio of the remaining number of iterations to the total number of iterations and the initial elimination ratio, wherein the initial elimination ratio is 20%.

[0187] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0188] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0189] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources, characterized in that the method... include: Adaptive optimization of the rolling time window is performed based on the predicted power grid operation status data and predicted electric vehicle access status data for the target area in the future time period to obtain the optimal rolling time window; The future time period is divided according to the optimal rolling time window to generate an adapted rolling time window sequence; Based on the adapted rolling time window sequence, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, the scheduling strategy for charging and discharging resources is optimized according to the preset monitoring feedback adjustment mechanism, and the optimal scheduling strategy is output for iterative regulation. This includes adaptive optimization of the scrolling time window to obtain the optimal scrolling time window, including: Load volatility is calculated based on the predicted power grid load sequence to obtain the power grid load volatility, wherein the power grid load volatility is the ratio of the standard deviation to the mean of the predicted power grid load in the predicted power grid load sequence; The load volatility is calculated based on the predicted vehicle charging load sequence to obtain the charging load volatility. The vehicle charging load share sequence is calculated based on the predicted power grid load sequence and the predicted vehicle charging load sequence, and the average load share and load share fluctuation are calculated. Adaptive optimization of the rolling time window is performed based on the grid load fluctuation, charging load fluctuation, average load ratio, and load ratio fluctuation to obtain the optimal rolling time window, including: The ratio of the preset standard power grid load fluctuation to the power grid load fluctuation is set as the first window compensation coefficient; The ratio of the preset standard charging load fluctuation to the charging load fluctuation is set as the second window compensation coefficient. The ratio of the preset standard load percentage average to the load percentage average is set as the third window compensation coefficient; The ratio of the preset standard load percentage fluctuation to the load percentage fluctuation is set as the fourth window compensation coefficient. The first window compensation coefficient, the second window compensation coefficient, the third window compensation coefficient, and the fourth window compensation coefficient are weighted and fused to generate a comprehensive window compensation coefficient; Based on the optimization and correction of the initial scrolling time window using the comprehensive window compensation coefficient, the optimal scrolling time window is output.

2. The rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources according to claim 1, characterized in that, Acquire predicted power grid operation status data and predicted electric vehicle access status data for the target area in the future time period, including: Obtain historical power grid load sequences and historical vehicle charging load sequences for the target area within a historical time period; Based on a long short-term memory network, and according to the historical power grid load sequence and the historical vehicle charging load sequence, a predicted power grid load sequence and a predicted vehicle charging load sequence for a future period are obtained as predicted power grid operating status data and predicted electric vehicle access status data, wherein the time interval of the future period is the same as the time interval of the historical period.

3. The rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources according to claim 2, characterized in that, Based on a Long Short-Term Memory (LSTM) network, and according to the historical power grid load sequence and historical vehicle charging load sequence, a predicted power grid load sequence and a predicted vehicle charging load sequence for future periods are obtained, including: Based on the historical power grid operation records of the target area, and constrained by the historical time period, a sample power grid load sequence set is collected, and the historical power grid load sequences of different sample power grid load sequences in the historical future time period are collected as sample predicted power grid load sequences to obtain a sample predicted power grid load sequence set. Using the sample power grid load sequence set as input and the sample predicted power grid load sequence set as supervision, a long short-term memory network is trained until convergence to generate a power grid load predictor. The power grid load forecaster is used to predict the power grid load sequence for future periods based on the historical power grid load sequence. Based on the historical power grid operation records of the target area, and constrained by the historical time period, a sample vehicle charging load sequence set and a sample predicted vehicle charging load sequence set are collected. Using the sample vehicle charging load sequence set as input and the sample predicted vehicle charging load sequence set as supervision, a long short-term memory network is trained until convergence to generate a vehicle charging load predictor. The vehicle charging load predictor uses the historical vehicle charging load sequence to predict the vehicle charging load sequence for future periods.

4. The rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources according to claim 3, characterized in that, Based on the adapted rolling time window sequence, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, the scheduling strategy for charging and discharging resources is optimized according to a preset monitoring feedback adjustment mechanism. The optimal scheduling strategy is then output for iterative regulation, including: In the sequence of adaptive scrolling time windows, the first adaptive scrolling time window is selected as the first adaptive scrolling time window, and the adjacent window of the first adaptive scrolling time window is selected as the second adaptive scrolling time window. The first historical power grid load sequence and the first historical vehicle charging load sequence within the first adaptation rolling time window are monitored and obtained. A short-term power grid load predictor and a short-term vehicle charging load predictor are constructed based on a long short-term memory network. A second predicted power grid load sequence and a second predicted vehicle charging load sequence are obtained based on the first historical power grid load sequence and the first historical vehicle charging load sequence. With the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, the scheduling strategy for charging and discharging resources is optimized according to a preset monitoring feedback adjustment mechanism, and the second optimal scheduling strategy is output. The second optimal scheduling strategy is used to schedule electric vehicle charging and discharging resources within the second adaptive rolling time window, and iterative regulation is performed based on the adaptive rolling time window sequence.

5. The rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources according to claim 4, characterized in that, With the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, the scheduling strategy for charging and discharging resources is optimized according to a preset monitoring feedback adjustment mechanism, and a second optimal scheduling strategy is output, including: Obtain the vehicle charging power adjustment space of the second adaptive rolling time window, and randomly select several initial charging powers within the vehicle charging power adjustment space. Based on the second predicted grid load sequence and the second predicted vehicle charging load sequence, grid load is predicted according to the plurality of initial charging powers, and a plurality of predicted grid load fluctuations are output. Based on the second predicted vehicle charging load sequence, vehicle charging simulation is performed according to the several initial charging powers, and several simulated vehicle charging efficiencies are output. With the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, several parameter fitnesss are calculated by weighting several predicted grid load fluctuations and several simulated vehicle charging efficiencies. The parameter fitnesss are negatively correlated with the predicted grid load fluctuations and positively correlated with the simulated vehicle charging efficiencies. Based on the aforementioned initial charging power and the aforementioned parameter fitness, the scheduling strategy for charging and discharging resources is optimized, and a second optimal scheduling strategy is output.

6. The rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources according to claim 5, characterized in that, Based on the aforementioned initial charging power and the aforementioned parameter fitness, an optimization strategy for scheduling charging and discharging resources is performed, and a second optimal scheduling strategy is output, including: Set the initial charging power as the initial solution, and arrange several initial solutions in descending order of parameter fitness to generate an initial solution sequence; The first solution in the initial solution sequence is selected as the optimal solution, and the remaining initial solutions are selected as the inferior solutions. The multiple inferior solutions are adjusted according to a preset step size, with the optimal solution as the direction, to obtain multiple updated inferior solutions. The optimal solutions and multiple updated inferior solutions are reordered according to the parameter fitness from large to small. After eliminating a preset proportion of the updated inferior solutions, an initial solution is randomly selected in the vehicle charging power adjustment space for equivalent supplementation to construct an updated solution sequence. Based on the updated solution sequence, iterative optimization continues until a preset number of convergences is reached. The optimal solution of the current updated solution sequence is then set as the second optimal scheduling strategy.

7. The rolling optimization method for dynamic scheduling of electric vehicle charging and discharging resources according to claim 6, characterized in that, The preset ratio is the product of the ratio of the remaining number of iterations to the total number of iterations and the initial elimination ratio, wherein the initial elimination ratio is 20%.

8. A rolling optimization dynamic scheduling system for electric vehicle charging and discharging resources, characterized in that, A method for implementing a rolling optimization dynamic scheduling method for electric vehicle charging and discharging resources as described in any one of claims 1-7 includes: The adaptive optimization module is used to adaptively optimize the rolling time window based on the predicted power grid operation status data and predicted electric vehicle access status data of the target area in the future time period, so as to obtain the optimal rolling time window. The sequence generation module is used to divide the future time period according to the optimal rolling time window and generate a sequence adapted to the rolling time window. The strategy optimization module is used to optimize the scheduling strategy of charging and discharging resources based on the adapted rolling time window sequence, with the goal of maximizing vehicle charging efficiency and minimizing grid load fluctuations, according to a preset monitoring feedback adjustment mechanism, and output the optimal scheduling strategy for iterative regulation. This includes adaptive optimization of the scrolling time window to obtain the optimal scrolling time window, including: Load volatility is calculated based on the predicted power grid load sequence to obtain the power grid load volatility, wherein the power grid load volatility is the ratio of the standard deviation to the mean of the predicted power grid load in the predicted power grid load sequence; The load volatility is calculated based on the predicted vehicle charging load sequence to obtain the charging load volatility. The vehicle charging load share sequence is calculated based on the predicted power grid load sequence and the predicted vehicle charging load sequence, and the average load share and load share fluctuation are calculated. Adaptive optimization of the rolling time window is performed based on the grid load fluctuation, charging load fluctuation, average load ratio, and load ratio fluctuation to obtain the optimal rolling time window, including: The ratio of the preset standard power grid load fluctuation to the power grid load fluctuation is set as the first window compensation coefficient; The ratio of the preset standard charging load fluctuation to the charging load fluctuation is set as the second window compensation coefficient. The ratio of the preset standard load percentage average to the load percentage average is set as the third window compensation coefficient; The ratio of the preset standard load percentage fluctuation to the load percentage fluctuation is set as the fourth window compensation coefficient. The first window compensation coefficient, the second window compensation coefficient, the third window compensation coefficient, and the fourth window compensation coefficient are weighted and fused to generate a comprehensive window compensation coefficient; Based on the optimization and correction of the initial scrolling time window using the comprehensive window compensation coefficient, the optimal scrolling time window is output.

Images