Smart micro-grid-based charging pile group load dynamic balance regulation method

Abstract

The application discloses a charging pile group load dynamic balance regulation method based on an intelligent microgrid, and relates to the technical field of load dynamic regulation.The application solves the technical problems of response lag caused by the disconnection between prediction and optimization decision and accelerated life attenuation caused by extensive energy storage scheduling.The application calculates the adjustable elastic coefficient of each vehicle by comprehensively considering the current state of charge, the minimum state of charge threshold, the available capacity of the battery, the discharge efficiency coefficient and the battery health degree attenuation factor, and dynamically divides the vehicles into the rigid regulation area, the semi-elastic regulation area or the high-elastic regulation area according to the adjustable elastic coefficient.In the power regulation process, the differential calling sequence of the semi-elastic area in the upward regulation and the high-elastic area in the downward regulation is used, so that the regulation pressure is matched with the regulation capacity of each vehicle and the influence degree on the user experience, the flexible regulation potential of the vehicle is fully released, and the influence on the user charging experience is reduced.

Description

Technical Field

[0001] This invention relates to the field of dynamic load regulation technology, specifically a method for dynamic load balancing regulation of charging pile groups based on smart microgrids. Background Technology

[0002] With the continuous growth of electric vehicle ownership, the disorderly charging behavior of large-scale electric vehicles poses a significant challenge to the safe and stable operation of the power distribution network. Smart microgrids, as a key technological framework integrating distributed photovoltaic (PV) power, energy storage systems, and charging loads, offer a feasible solution to these problems. However, the intermittency and volatility of PV power output, the randomness and peak-valley characteristics of charging loads, and the unpredictability of user charging behavior in microgrids combine to create significant difficulties in achieving real-time power balance among power sources, grids, loads, and storage.

[0003] In existing technologies, various solutions have been proposed for load regulation of charging pile groups. Some solutions determine the control commands of each module by obtaining the grid connection point operation curve, but relying on preset operation curves results in insufficient responsiveness to real-time dynamic changes. Some solutions adjust the output power of charging piles by dividing charging speed intervals, but their rule-based static division logic lacks the ability to predict and anticipate changes in operating conditions. Other solutions propose a method of segmenting the remaining charging tasks and labeling them with flexible tags, but their power adjustment strategy is still essentially based on fixed rules, making it difficult to achieve global optimization when faced with multiple conflicting objectives. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a dynamic load balancing adjustment method for charging pile groups based on smart microgrids, which solves the problems of delayed response caused by the separation of prediction and optimization decision-making, and accelerated lifespan degradation caused by extensive energy storage scheduling.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a dynamic load balancing adjustment method for charging pile groups based on a smart microgrid, which specifically includes the following steps: The photovoltaic prediction curve and charging load prediction curve of the charging pile group are obtained based on the long short-term memory neural network model. The charging and discharging behavior constraint parameter set of electric vehicles is constructed, the cumulative deviation of the ultra-short-term prediction sequence and the day-ahead scheduling plan is calculated, and dynamic optimization or power fine-tuning is required. For dynamic optimization, the adjustable elasticity coefficient of each electric vehicle is calculated and different adjustment zones are divided. At the same time, the net power adjustment requirement is calculated, and the charging power is adjusted upward or downward accordingly. The process of increasing charging power includes first allocating the power from the semi-flexible adjustment zone according to the proportion of the adjustable capacity. If the power is insufficient, it is then allocated from the high-flexibility adjustment zone according to the proportion of the adjustable capacity. If there is still excess power after increasing to the rated power, some vehicles in the high-flexibility adjustment zone will be switched to discharge mode. The process of reducing charging power includes allocating it from the high-flexibility adjustment zone according to the proportion of the adjustable capacity and gradually reducing it to zero. If it is insufficient, it is then allocated from the semi-flexible adjustment zone according to the proportion of the adjustable capacity and reduced to the minimum value allowed for each vehicle. The power fine-tuning process includes calculating the cumulative power deviation within a preset time window. When the cumulative power deviation is greater than the dead zone threshold but not greater than the trigger threshold, the net adjustment amount is calculated simultaneously. The energy storage system is given priority in handling the net adjustment amount. If the energy storage system is unavailable, vehicles are selected from the high-elasticity adjustment zone and the net adjustment amount is allocated according to the elasticity coefficient ratio.

[0006] As a further aspect of the present invention, the method for determining whether dynamic optimization processing or power fine-tuning processing is required is as follows: The system acquires historical charging load statistics for the same period of the charging pile cluster and the weather forecast data for the next day. The acquired data is then input into a pre-built long short-term memory neural network model to obtain the photovoltaic prediction curve and charging load prediction curve for each time period of the next day. Based on the historical operating data, real-time weather data and current charging connection status within the sliding window 15 minutes ahead of the current time, the system generates a future short-term charging load prediction sequence and a photovoltaic output prediction sequence. Based on the obtained ultra-short-term forecast sequence and the day-ahead scheduling plan, the deviation between the two is calculated hourly, and the cumulative deviation is calculated. At the same time, it is compared with the preset deviation threshold. If the cumulative deviation exceeds the preset deviation threshold, dynamic optimization is required; otherwise, power fine-tuning is required.

[0007] As a further aspect of the present invention, the method for calculating the adjustable elastic coefficient of each electric vehicle and dividing it into different adjustment zones is as follows: According to the formula ,in This indicates the current percentage of battery charge in vehicle i. This indicates the minimum percentage of charge in vehicle battery i. Indicates the available battery capacity. This represents the discharge efficiency coefficient. This represents the battery health degradation factor. Based on the obtained adjustable elastic coefficient Ei, electric vehicles are divided into different adjustment zones. If the adjustable elastic coefficient Ei = 0, they are classified into the rigid adjustment zone. If the adjustable elastic coefficient Ei is greater than 0, they are classified into the semi-elastic adjustment zone. If the adjustable elastic coefficient Ei is greater than the set discharge threshold, they are classified into the high elastic adjustment zone.

[0008] As a further aspect of the present invention, the method for calculating the net power adjustment requirement is as follows: According to the formula Calculated net power adjustment requirements ,in For charging load deviation, For photovoltaic power output deviation, To account for external power grid switching deviations, and to adjust according to net power requirements. The sign of the net power adjustment requirement determines the direction of the call. If the net power adjustment requirement is negative, the charging power needs to be increased; conversely, if the net power adjustment requirement is positive, the charging power needs to be decreased.

[0009] As a further aspect of the present invention, the method of switching a portion of the vehicles in the high-elasticity adjustment zone to discharge mode is as follows: Discharge candidate vehicles are selected based on preset judgment conditions and sorted from largest to smallest according to their corresponding elasticity coefficients. The maximum allowable discharge power of the discharge candidate vehicles in the sorted list is gradually accumulated until the accumulated value meets the net power adjustment requirements. If the accumulated maximum discharge power of all discharge candidate vehicles still cannot meet the net power adjustment requirements, all discharge candidate vehicles will participate in the discharge according to the maximum allowable discharge power, and the missing part will be supplemented by the power grid.

[0010] As a further aspect of the present invention, the preset determination conditions include: The current state of charge is at least 10% higher than its minimum state of charge threshold; the estimated battery health value is greater than the preset degradation threshold; the vehicle and the charging station support bidirectional charging and discharging. Based on the judgment criteria, electric vehicles that do not meet any of the criteria are eliminated, and the remaining electric vehicles are marked as discharge candidate vehicles.

[0011] As a further aspect of the present invention, the power fine-tuning process includes: The ultra-short-term forecast is compared point by point with the day-ahead scheduling plan to calculate the cumulative power deviation within the next 15-minute time window. At the same time, the cumulative power deviation is compared with the preset trigger threshold. The preset trigger threshold is automatically calculated based on the transformer's rated capacity and the statistical distribution of historical forecast errors. When the cumulative power deviation is less than the dead zone threshold, the current power command of each device remains unchanged, and only continuous monitoring is performed without triggering any adjustments. When the cumulative power deviation is greater than the trigger threshold, no power fine-tuning is performed, and dynamic optimization is directly carried out.

[0012] As a further aspect of the present invention, when the cumulative power deviation is greater than the dead zone threshold and less than the trigger threshold, the difference between the cumulative power deviation and the dead zone threshold is calculated to obtain the net adjustment amount. If the current state of charge of the energy storage system is between 20% and 90%, the energy storage system is given priority to undertake fine-tuning. Specifically, the net adjustment amount is directly superimposed on the current charging and discharging power command of the energy storage system, and transitions to the new command according to its own maximum allowable power change rate.

[0013] As a further aspect of the present invention, if the energy storage system is unavailable, a number of vehicles with the largest elasticity coefficients are selected from the high elasticity adjustment zone to undertake fine-tuning. The net adjustment amount is allocated proportionally according to the elasticity coefficients of the selected vehicles. The fine-tuning command is monitored. If the transformer load rate exceeds 85% or any phase current exceeds 80% of the rated value, the fine-tuning is stopped. If the energy storage charge state reaches the 20% or 90% boundary during the fine-tuning process, the fine-tuning workload of the energy storage will be automatically transferred to the high-elasticity vehicle. When the cumulative power deviation is detected to fall back to within the dead zone threshold and remain stable for more than 2 control cycles, the correction amount generated by the fine-tuning will be gradually reduced to zero in a ramp-down manner.

[0014] This invention provides a method for dynamic load balancing adjustment of charging pile groups based on smart microgrids. Compared with existing technologies, it has the following advantages: This invention constructs a set of charging and discharging behavior constraint parameters that reflect the individual behavioral characteristics of each vehicle by statistically extracting the charging period, average dwell time, and historical state of charge changes from the historical charging records of each electric vehicle. This provides a data-driven objective basis for subsequent calculation of adjustable elastic coefficients and division of adjustment zones, overcoming the deficiency in existing technologies that lack quantitative description of individual vehicle behavior differences.

[0015] This invention calculates the adjustable elasticity coefficient for each vehicle by comprehensively considering the current state of charge, minimum state of charge threshold, available battery capacity, discharge efficiency coefficient, and battery health degradation factor. Based on this, the vehicle is dynamically divided into a rigid adjustment zone, a semi-elastic adjustment zone, or a high-elasticity adjustment zone. During power adjustment, the differentiated calling order prioritizes the semi-elastic zone when adjusting upwards and the high-elastic zone when adjusting downwards, so that the adjustment pressure matches the adjustment capability of each vehicle and the degree of impact on user experience. This fully releases the flexible adjustment potential of the vehicle while reducing the impact on the user's charging experience.

[0016] In the power fine-tuning process, this invention prioritizes the use of the energy storage system. In the dynamic optimization process, it calls vehicle resources in different adjustment zones in stages according to the positive and negative directions of the net power adjustment demand. When the vehicle discharge is insufficient, the grid supplements it, and when the energy storage exceeds the limit, the vehicle takes over. This forms a complementary advantage between the rapid response of energy storage and the deep adjustment of vehicles, overcoming the defect that energy storage and vehicle scheduling are independent in the prior art. Attached Figure Description

[0017] Figure 1 This is a flowchart of the dynamic load balance adjustment method for charging pile groups according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] First Embodiment Please see Figure 1 This application provides a method for dynamic load balancing adjustment of charging pile groups based on smart microgrids, which specifically includes the following steps: The photovoltaic prediction curve and charging load prediction curve of the charging pile group are obtained based on the long short-term memory neural network model. The charging and discharging behavior constraint parameter set of electric vehicles is constructed, the cumulative deviation of the ultra-short-term prediction sequence and the day-ahead scheduling plan is calculated, and dynamic optimization or power fine-tuning is required. For dynamic optimization, the adjustable elasticity coefficient of each electric vehicle is calculated and different adjustment zones are divided. At the same time, the net power adjustment requirement is calculated, and the charging power is adjusted upward or downward accordingly. The process of increasing charging power includes first allocating the power from the semi-flexible adjustment zone according to the proportion of the adjustable capacity. If the power is insufficient, it is then allocated from the high-flexibility adjustment zone according to the proportion of the adjustable capacity. If there is still excess power after increasing to the rated power, some vehicles in the high-flexibility adjustment zone will be switched to discharge mode. The process of reducing charging power includes allocating it from the high-flexibility adjustment zone according to the proportion of the adjustable capacity and gradually reducing it to zero. If it is insufficient, it is then allocated from the semi-flexible adjustment zone according to the proportion of the adjustable capacity and reduced to the minimum value allowed for each vehicle. The power fine-tuning process includes calculating the cumulative power deviation within a preset time window. When the cumulative power deviation is greater than the dead zone threshold but not greater than the trigger threshold, the net adjustment amount is calculated. The energy storage system is given priority in undertaking the net adjustment amount. If the energy storage system is unavailable, vehicles are selected from the high-elasticity adjustment zone and the net adjustment amount is allocated according to the elasticity coefficient ratio.

[0020] Second Embodiment As a second embodiment of the present invention, it is implemented based on the first embodiment, and the difference from the first embodiment is as follows: The system acquires historical charging load statistics for the same period of the charging pile cluster and the weather forecast data for the next day. The acquired data is then input into a pre-constructed long short-term memory neural network model. The input layer of the long short-term memory neural network model is multi-dimensional time series data, specifically including historical photovoltaic data and weather forecast data. The encoding layer includes a first LSTM layer and a second LSTM layer. Both LSTM layers use the tanh activation function to control the state update of the control unit. After the second LSTM layer, a fully connected layer is connected to map the 64-dimensional compressed features to the predicted output dimension. The output layer includes 24 neurons, corresponding to the hourly photovoltaic output of the 24 hours of the prediction day. This yields the photovoltaic prediction curve and charging load prediction curve for each time period of the next day. At the same time, the system acquires the historical charging records of each electric vehicle, extracts the charging time period, average dwell time, and historical state of charge changes of each electric vehicle, and then constructs a corresponding set of charging and discharging behavior constraint parameters based on the obtained electric vehicle parameters. Next, the real-time power, voltage, current, connection status and vehicle charge status of each charging gun are collected, the charge status and charging and discharging power of each energy storage unit are collected, the power generation and irradiance of the photovoltaic inverter are collected, and based on the historical operating data, real-time meteorological data and current charging connection status within the 15-minute sliding window before the current time, a future short-term charging load prediction sequence and a photovoltaic output prediction sequence are generated. Based on the obtained ultra-short-term forecast sequence and the day-ahead scheduling plan, which is set by the operator, the deviation between the two is calculated hourly, and the cumulative deviation is calculated. At the same time, it is compared with the preset deviation threshold, and the specific value of the preset deviation threshold is set by the operator. If the cumulative deviation exceeds the preset deviation threshold, dynamic optimization is required; otherwise, power fine-tuning is required.

[0021] For situations requiring dynamic optimization, the charging and discharging behavior constraint parameter set for all electric vehicles currently charging is obtained. Combined with the current measured battery level, user-set time, and target battery level, the adjustable elasticity coefficient Ei for each electric vehicle is calculated. Specifically... ,in This indicates the current percentage of battery charge in vehicle i. This indicates the minimum percentage of charge in vehicle battery i. Indicates the available battery capacity. This represents the discharge efficiency coefficient. This represents the battery health degradation factor. Based on the obtained adjustable elastic coefficient Ei, electric vehicles are divided into different adjustment zones. If the adjustable elastic coefficient Ei = 0, they are classified into the rigid adjustment zone. If the adjustable elastic coefficient Ei is greater than 0, they are classified into the semi-elastic adjustment zone. If the adjustable elastic coefficient Ei is greater than the set discharge threshold, they are classified into the high elastic adjustment zone. Next, obtain the current transformer load rate, energy storage state of charge, and photovoltaic output changes, and then apply the formula... Calculated net power adjustment requirements ,in For charging load deviation, For photovoltaic power output deviation, To account for external power grid switching deviations, and to adjust according to net power requirements. The sign of the net power adjustment requirement determines the direction of the call. If the net power adjustment requirement is negative, the charging power needs to be increased; conversely, if the net power adjustment requirement is positive, the charging power needs to be decreased. In response to the situation of increasing the charging power, the net power adjustment demand is first allocated from the semi-flexible adjustment zone according to the proportion of the adjustable capacity of electric vehicles. If the net power adjustment demand is still not met, the net power adjustment demand is then allocated from the high-flexibility adjustment zone according to the proportion of the adjustable capacity of each electric vehicle. If there is still an excess after increasing to the rated power of each electric vehicle, some vehicles in the high-flexibility adjustment zone will be switched to discharge mode. In the event of a reduction in charging power, the net power adjustment requirement is allocated from the high-elasticity adjustment zone according to the proportion of the capacity that each electric vehicle can reduce, and its charging power is gradually reduced to zero. If the net power adjustment requirement is still not met, the net power adjustment requirement is then allocated from the semi-elasticity adjustment zone according to the proportion of the capacity that each electric vehicle can reduce, and its charging power is reduced to the minimum allowable value for each electric vehicle.

[0022] Third Embodiment As a third embodiment of the present invention, it is implemented based on the second embodiment, and the difference from the second embodiment is as follows: The specific method for switching some vehicles in the high-elasticity adjustment zone to discharge mode is as follows: discharge vehicles are screened based on preset judgment conditions, which include: the current state of charge is at least 10% higher than its lowest state of charge threshold; the estimated battery health value is greater than the preset degradation threshold; the vehicle and the charging pile support bidirectional charging and discharging; based on the above judgment conditions, electric vehicles that do not meet any of the conditions are eliminated, and the remaining electric vehicles are marked as discharge candidate vehicles. Obtain candidate vehicles for discharge and sort them from largest to smallest according to their corresponding elasticity coefficients. Based on the candidate vehicles in the sorted list, gradually accumulate their maximum allowable discharge power until the accumulated value meets the net power adjustment requirements. If the accumulated maximum discharge power of all candidate vehicles still cannot meet the net power adjustment requirements, then all candidate vehicles will participate in discharge according to their maximum allowable discharge power, and the missing part will be supplemented by the power grid. During the reverse transition of the power command, a transition time window is set, and the transition time = |target discharge power - current charging power| / maximum allowable power change rate limit. The maximum allowable power change rate limit is the product of the maximum current change rate specified by the battery management system and the current DC bus voltage. Within the transition time window, the power command is reduced from the current charging power value to zero in a ramp manner, and then continues to be reduced to the target discharge power value in a ramp manner, and the ramp slope does not exceed the maximum current change rate specified by the battery management system. At the same time, the vehicle discharge process is monitored in real time. When the vehicle enters steady-state discharge, the actual discharge power is calculated based on the product of DC side voltage and current, and the relative deviation between it and the commanded discharge power is calculated. The relative deviation is compared with the allowable error range. The allowable error range is automatically determined according to the power control accuracy level of the charging pile terminal. For terminals with an accuracy level of 1, the error range is ±5%; for terminals with an accuracy level of 0.5, the error range is ±2%. If it is not within the error range, the discharge mode is automatically exited. For situations requiring power fine-tuning, the ultra-short-term forecast is compared point-by-point with the day-ahead scheduling plan to calculate the cumulative power deviation within the next 15-minute time window. Simultaneously, the cumulative power deviation is compared with a preset trigger threshold. The preset trigger threshold is automatically calculated based on the transformer's rated capacity and the statistical distribution of historical forecast errors, and is set to 5% of the transformer's rated power. The dead zone threshold is set to one-quarter of the trigger threshold. When the cumulative power deviation is less than the dead zone threshold, the current power commands for each device remain unchanged, and only continuous monitoring is performed without triggering any adjustments. When the cumulative power deviation is greater than the trigger threshold, power fine-tuning is not performed, and dynamic optimization is directly implemented. When the cumulative power deviation is greater than the dead zone threshold and less than the trigger threshold, the difference between the cumulative power deviation and the dead zone threshold is calculated to obtain the net adjustment amount. If the current state of charge of the energy storage system is between 20% and 90%, the energy storage system is given priority to undertake fine-tuning. Specifically, the net adjustment amount is directly superimposed on the current charging and discharging power command of the energy storage system, and the system transitions to the new command according to its maximum allowable power change rate. If the energy storage system is unavailable, several vehicles with the largest elasticity coefficients from the high-elasticity adjustment zone are selected to undertake fine-tuning. The net adjustment amount is allocated proportionally according to the elasticity coefficients of the selected vehicles. The fine-tuning commands are monitored. If the transformer load rate exceeds 85% or any phase current exceeds 80% of the rated value, the fine-tuning is stopped. If the energy storage charge state touches the 20% or 90% boundary during the fine-tuning process, the fine-tuning amount of the energy storage is automatically transferred to the high-elasticity vehicles. When the cumulative power deviation is detected to fall back to within the dead zone threshold and remain stable for more than 2 control cycles, the correction amount generated by the fine-tuning is gradually reduced to zero in a ramp-down manner.

[0023] Fourth embodiment As a fourth embodiment of the present invention, the key point is that the implementation process of all the above embodiments is combined and implemented.

[0024] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0025] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for dynamic load balancing adjustment of charging pile groups based on smart microgrids, characterized in that, The method specifically includes the following steps: The photovoltaic prediction curve and charging load prediction curve of the charging pile group are obtained based on the long short-term memory neural network model. The charging and discharging behavior constraint parameter set of electric vehicles is constructed, the cumulative deviation of the ultra-short-term prediction sequence and the day-ahead scheduling plan is calculated, and dynamic optimization or power fine-tuning is required. For dynamic optimization, the adjustable elasticity coefficients of each electric vehicle are calculated and different adjustment zones are defined. The specific processing method is as follows: According to the formula ,in This indicates the current percentage of battery charge in vehicle i. This indicates the minimum percentage of charge in vehicle battery i. Indicates the available battery capacity. This represents the discharge efficiency coefficient. This represents the battery health degradation factor, based on the obtained adjustable elasticity coefficient E. i Electric vehicles are divided into different adjustment zones, with an adjustable elasticity coefficient E. i =0, then it is classified into the rigid adjustment zone, if the adjustable elastic coefficient E i If the value is greater than 0, it is classified into the semi-elastic adjustment zone. If the adjustable elastic coefficient E i If the discharge threshold is exceeded, it is classified into the high-elasticity adjustment zone. Simultaneously, the net power adjustment requirement is calculated, and the charging power is adjusted upwards or downwards accordingly. The specific processing method is as follows: According to the formula Calculated net power adjustment requirements ,in This is due to charging load deviation. For photovoltaic power output deviation, To account for external power grid switching deviations, and to adjust according to net power requirements. The sign of the net power adjustment requirement determines the direction of the call. If the net power adjustment requirement is negative, the charging power needs to be increased; conversely, if the net power adjustment requirement is positive, the charging power needs to be decreased. The process of increasing charging power includes first allocating the power from the semi-flexible adjustment zone according to the proportion of the adjustable capacity. If the power is insufficient, it is then allocated from the high-flexibility adjustment zone according to the proportion of the adjustable capacity. If there is still excess power after increasing to the rated power, some vehicles in the high-flexibility adjustment zone will be switched to discharge mode. The process of reducing charging power includes allocating it from the high-flexibility adjustment zone according to the proportion of the adjustable capacity and gradually reducing it to zero. If it is insufficient, it is then allocated from the semi-flexible adjustment zone according to the proportion of the adjustable capacity and reduced to the minimum value allowed for each vehicle. The power fine-tuning process includes calculating the cumulative power deviation within a preset time window. When the cumulative power deviation is greater than the dead zone threshold but not greater than the trigger threshold, the net adjustment amount is calculated simultaneously. The energy storage system is given priority in handling the net adjustment amount. If the energy storage system is unavailable, vehicles are selected from the high-elasticity adjustment zone and the net adjustment amount is allocated according to the elasticity coefficient ratio.

2. The method for dynamic load balancing adjustment of charging pile groups based on smart microgrids according to claim 1, characterized in that, The method for determining whether dynamic optimization or power fine-tuning is required is as follows: The system acquires historical charging load statistics for the same period of the charging pile group and the weather forecast data for the next day. The acquired data is then input into a pre-built long short-term memory neural network model to obtain the photovoltaic prediction curve and charging load prediction curve for each time period of the next day. Based on the historical operating data, real-time weather data and current charging connection status within a 15-minute sliding window from the current time, an ultra-short-term prediction sequence is generated, which includes a future short-term charging load prediction sequence and a photovoltaic output prediction sequence. Based on the obtained ultra-short-term forecast sequence and the day-ahead scheduling plan, the deviation between the two is calculated hourly, and the cumulative deviation is calculated. At the same time, it is compared with the preset deviation threshold. If the cumulative deviation exceeds the preset deviation threshold, dynamic optimization is required; otherwise, power fine-tuning is required.

3. The method for dynamic load balancing adjustment of charging pile groups based on smart microgrids according to claim 1, characterized in that, The method for switching some vehicles in the high-elasticity adjustment zone to discharge mode is as follows: Discharge candidate vehicles are selected based on preset judgment conditions and sorted from largest to smallest according to their corresponding elasticity coefficients. The maximum allowable discharge power of the discharge candidate vehicles in the sorted list is gradually accumulated until the accumulated value meets the net power adjustment requirements. If the accumulated maximum discharge power of all discharge candidate vehicles still cannot meet the net power adjustment requirements, all discharge candidate vehicles will participate in the discharge according to the maximum allowable discharge power, and the missing part will be supplemented by the power grid.

4. The method for dynamic load balancing adjustment of charging pile groups based on smart microgrids according to claim 3, characterized in that, The preset judgment conditions include: The current state of charge is at least 10% higher than its minimum state of charge threshold; the estimated battery health value is greater than the preset degradation threshold; the vehicle and the charging station support bidirectional charging and discharging. Based on the judgment criteria, electric vehicles that do not meet any of the criteria are eliminated, and the remaining electric vehicles are marked as discharge candidate vehicles.

5. The method for dynamic load balancing adjustment of charging pile groups based on smart microgrids according to claim 1, characterized in that, The power fine-tuning process includes: The ultra-short-term forecast sequence is compared point by point with the day-ahead scheduling plan to calculate the cumulative power deviation within the next 15-minute time window. At the same time, the cumulative power deviation is compared with the preset trigger threshold. The preset trigger threshold is automatically calculated based on the transformer's rated capacity and the statistical distribution of historical forecast errors. When the cumulative power deviation is less than the dead zone threshold, the current power command of each device remains unchanged, and only continuous monitoring is performed without triggering any adjustments. When the cumulative power deviation is greater than the trigger threshold, no power fine-tuning is performed, and dynamic optimization is directly carried out.

6. The method for dynamic load balancing adjustment of charging pile groups based on smart microgrids according to claim 5, characterized in that, When the cumulative power deviation is greater than the dead zone threshold but less than the trigger threshold, the difference between the cumulative power deviation and the dead zone threshold is calculated to obtain the net adjustment amount. If the current state of charge of the energy storage system is between 20% and 90%, the energy storage system is used to carry out fine-tuning. Specifically, the net adjustment amount is directly superimposed on the current charging and discharging power command of the energy storage system, and the system transitions to the new command according to its maximum allowable power change rate.

7. The method for dynamic load balancing adjustment of charging pile groups based on smart microgrids according to claim 6, characterized in that, If the energy storage system is unavailable, select several vehicles with the largest elasticity coefficient from the high elasticity adjustment zone to undertake fine-tuning. Allocate the net adjustment amount proportionally according to the elasticity coefficient of the selected vehicles. Monitor the fine-tuning instructions. If the transformer load rate exceeds 85% or any phase current exceeds 80% of the rated value, the fine-tuning is stopped. If the energy storage charge state reaches the 20% or 90% boundary during the fine-tuning process, the fine-tuning workload of the energy storage will be automatically transferred to the high-elasticity vehicle. When the cumulative power deviation is detected to fall back to within the dead zone threshold and remain stable for more than 2 control cycles, the correction amount generated by the fine-tuning will be gradually reduced to zero in a ramp-down manner.

Images