Dynamic load balancing power supply system for new energy automobile charging pile cluster

Abstract

The invention, which relates to the technical field of the new energy vehicle power supply system, discloses a dynamic load balancing power supply system for a new energy vehicle charging pile cluster, comprising a data acquisition module, a load prediction and decision module, a dynamic power distribution execution module and a safety protection and communication module. The data acquisition module is used for monitoring three-phase electrical parameters, temperature and electric energy quality data of the charging pile cluster in real time; the load prediction and decision module is connected with the data acquisition module and is used for predicting a load trend based on historical and real-time data and generating a dynamic regulation and control strategy; the dynamic power distribution execution module is connected with the load prediction and decision module, accurate prediction and rapid dynamic regulation and control of the charging load can be achieved, the problems of three-phase imbalance, transformer loss, harmonic pollution and the like of a power distribution system caused by randomness and volatility of charging behaviors are effectively solved, and the intelligent graded safety protection capability is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of new energy vehicle power supply system, in particular to a dynamic load balancing power supply system for new energy vehicle charging pile cluster. BACKGROUND

[0002] As a key link to support the large-scale application of electric vehicles, charging pile cluster is evolving towards high power, high density and centralization. Its operation characteristics significantly increase the power supply pressure of the local distribution network. The randomness and volatility of charging behavior change the charging pile cluster from a passive load to an active and unpredictable disturbance source, posing new challenges to the power quality and safe and stable operation of the regional power grid.

[0003] Currently, the system for supplying power to the charging pile cluster usually adopts fixed power distribution or simple round-robin control strategy. Although such conventional power supply system can realize basic functions, it often responds slowly and has limited regulation capacity when dealing with dynamic scenarios such as peak period concentrated charging and vehicle random start-stop.

[0004] The deficiencies of the prior art mainly lie in the following aspects: the system lacks foresight and fast dynamic regulation capacity for three-phase load imbalance, which easily leads to increased transformer loss, neutral line overcurrent and even equipment overheating; it lacks effective real-time monitoring and compensation means for power quality problems such as harmonics generated during charging, affecting the power supply quality of the power grid; the communication architecture lacks redundancy and the safety protection mechanism is relatively single, making it difficult to achieve intelligent response of grading and self-healing under abnormal conditions. Therefore, we propose a dynamic load balancing power supply system for new energy vehicle charging pile cluster. SUMMARY

[0005] The technical problem to be solved by the present application is to overcome the existing defects and provide a dynamic load balancing power supply system for new energy vehicle charging pile cluster, which can realize accurate prediction and fast dynamic regulation of charging load, effectively solve the problems of three-phase imbalance, transformer loss and harmonic pollution of the distribution system caused by the randomness and volatility of charging behavior, and has intelligent grading safety protection capability, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, this invention provides the following technical solution: a dynamic load balancing power supply system for new energy vehicle charging pile clusters, comprising a data acquisition module, a load prediction and decision module, a dynamic power distribution execution module, and a safety protection and communication module. The data acquisition module is used to monitor the three-phase electrical parameters, temperature, and power quality data of the charging pile cluster in real time. The load prediction and decision module is connected to the data acquisition module and is used to predict load trends based on historical and real-time data and generate dynamic control strategies. The dynamic power distribution execution module is connected to the load prediction and decision module and is used to quickly control power distribution according to the control strategies to achieve three-phase load balancing. The safety protection and communication module is connected to the data acquisition module, the load prediction and decision module, and the dynamic power distribution execution module respectively, and is used to provide a multi-level safety response mechanism and ensure reliable communication between modules within the system. By constructing a complete closed-loop control system that includes four major functions—sensing, decision-making, execution, and protection—this system systematically solves problems such as three-phase imbalance, transformer overload, and harmonic pollution in the power distribution system caused by random start-stop and power fluctuations in charging pile clusters, thereby improving the safety and stability of the power grid and optimizing charging efficiency.

[0007] Furthermore, the data acquisition module includes a multimodal sensing unit, a synchronous acquisition engine, and a harmonic analysis unit. The multimodal sensing unit is configured at each charging pile power supply node to collect three-phase current, voltage, power factor, and equipment temperature parameters. The synchronous acquisition engine is used to achieve data clock synchronization across power supply nodes. The harmonic analysis unit is used to analyze the harmonic components in the current and voltage waveforms in real time. By setting up the multimodal sensing unit, synchronous acquisition engine, and harmonic analysis unit, a comprehensive data foundation with high accuracy, high synchronization, and power quality details is provided for subsequent load prediction and decision-making. This ensures the accuracy and real-time nature of the information on which the system decisions depend, creating conditions for precise control.

[0008] Furthermore, the load forecasting and decision-making module includes a hybrid forecasting model and an edge computing controller. The hybrid forecasting model is composed of an LSTM neural network and an ARIMA algorithm fused with dynamic weights to generate load forecast curves for the next 15-30 minutes. The edge computing controller is used to run the hybrid forecasting model and execute the model predictive control algorithm to generate a real-time power distribution control strategy that includes relay timing parameters and capacitor compensation schemes. By adopting the hybrid forecasting model and the edge computing controller, accurate short-term forecasts of charging load trends can be achieved, and the optimal control strategy can be quickly generated at the edge side near the data source based on the forecast results, thereby overcoming the problem of response lag in traditional solutions and realizing forward-looking dynamic power regulation.

[0009] Furthermore, the dynamic power distribution execution module includes a three-phase balance control unit and a harmonic compensation unit. The three-phase balance control unit is used to control the solid-state relay array to realize the dynamic migration of the charging pile between different phase lines according to the strategy generated by the edge computing controller. The harmonic compensation unit is used to control the switching of the parallel compensation capacitor bank according to the harmonic analysis results. By setting up the three-phase balance control unit and the harmonic compensation unit, the control strategy can be executed directly and quickly, specifically realizing the dynamic switching of the charging pile phase lines to balance the three-phase load, and at the same time compensating for grid harmonics, thereby effectively reducing transformer losses and improving power quality.

[0010] Furthermore, the safety protection and communication module includes a multi-level safety response unit and a multi-mode communication unit. The multi-level safety response unit is used to implement gradient cooling control, charging priority screening, or self-healing grid reconfiguration based on the severity of the anomaly. The multi-mode communication unit integrates power line carrier communication and wireless communication links for reliable data transmission within the system and with the upper-level platform. By setting up multi-level safety response units and multi-mode communication units, it is ensured that the system can take differentiated protection measures according to the severity of the situation when abnormal operating conditions such as over-temperature, overload, or equipment failure occur. The redundant communication links ensure the reliable transmission of control commands, greatly enhancing the robustness and self-healing capability of the system.

[0011] Furthermore, the dynamic weight allocation rule in the hybrid prediction model is as follows: when the system detects that the load fluctuation rate is greater than a set threshold, the weight ratio of the LSTM neural network is automatically increased; when the load change is stable, the weight ratio of the ARIMA algorithm is increased. By specifying the dynamic weight allocation rule of the hybrid prediction model, the prediction model can adapt to the dynamic change characteristics of the load. When the load changes abruptly, the LSTM neural network is used first to handle nonlinear characteristics, and when the load is stable, the ARIMA algorithm is used first to capture periodic patterns, thereby maintaining high prediction accuracy under various operating conditions.

[0012] Furthermore, the constraints for the three-phase balance control unit to perform phase migration include a source phase line load deviation of more than 15%, a target phase line remaining capacity greater than 120% of the power demand of the charging pile to be migrated, and the number of charging piles migrated in a single batch not exceeding 20% ​​of the total number of the cluster. By setting specific constraints for phase migration, the three-phase balance control unit is guided to make decisions on the premise of ensuring the safe and effective migration operation, avoiding new imbalances or equipment overload caused by the migration operation itself, and achieving a stable and controllable load balancing process.

[0013] Furthermore, the charging priority screening mechanism in the multi-level safety response unit generates an interruption sequence based on the user's historical credit score, the vehicle's charging urgency, and the battery's health status. By clarifying the judgment dimensions for charging priority screening, it can fairly, reasonably, and intelligently determine the order of suspending charging services when the system must reduce its load, prioritizing the emergency charging needs of important users while also taking into account battery health, thereby improving user satisfaction while ensuring system safety.

[0014] Furthermore, when the multi-mode communication unit detects that the quality of the primary communication channel is lower than a threshold, it can automatically switch to the backup wireless communication link. By specifying the automatic switching function of the communication link, the system can seamlessly switch to the backup wireless link when the quality of the primary communication channel (such as power line carrier) deteriorates due to interference or other reasons, ensuring that critical monitoring data and control commands are not interrupted and maintaining the coordinated operation between various modules of the system.

[0015] Furthermore, it also includes an interface with the virtual power plant management cloud platform, used to receive control instructions from the load aggregation platform and report system operation data. By setting up an interface with the virtual power plant platform, the local charging pile cluster can be used as a flexible and adjustable aggregated resource to participate in broader grid dispatch and demand response, thereby improving the economy and reliability of the entire power system and expanding the application value of the system.

[0016] Compared with the prior art, the beneficial effects of the present invention are as follows: This dynamic load balancing power supply system for new energy vehicle charging pile clusters has the following advantages: 1. By introducing multimodal sensing and synchronous acquisition technology, and combining it with real-time harmonic analysis, the system has built a high-precision and highly synchronous global electrical parameter sensing capability, providing a reliable data foundation for subsequent intelligent decision-making and precise control, and improving the system's insight and response speed to power quality problems from the source.

[0017] 2. By adopting a dynamic weighted fusion prediction model based on LSTM and ARIMA, the system can adapt to the stable and sudden characteristics of the load, thereby achieving high-precision short-term load prediction under various operating conditions. This provides a core decision basis for forward-looking dynamic power regulation and three-phase balance optimization, effectively overcoming the drawbacks of slow response of traditional methods.

[0018] 3. A comprehensive protection mechanism integrating multi-level security response and redundant communication was designed. The system can intelligently implement differentiated protection strategies from gradient load reduction to network reconstruction according to the severity of the anomaly. The system also ensures the absolute reliability of control commands through automatic and seamless switching of communication links, which significantly enhances the robustness and self-healing ability of the system in complex operating environments. Attached Figure Description

[0019] Figure 1 This is a flowchart of the workflow of the present invention. Detailed Implementation

[0020] 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.

[0021] Please see Figure 1 This embodiment provides a technical solution: a dynamic load balancing power supply system for new energy vehicle charging pile clusters, including a data acquisition module, a load prediction and decision module, a dynamic power distribution execution module, and a safety protection and communication module. The data acquisition module is used to monitor the three-phase electrical parameters, temperature, and power quality data of the charging pile cluster in real time. The load prediction and decision module is connected to the data acquisition module and is used to predict load trends based on historical and real-time data and generate dynamic control strategies. The dynamic power distribution execution module is connected to the load prediction and decision module and is used to quickly control power distribution according to the control strategy to achieve three-phase load balancing. The safety protection and communication module is connected to the data acquisition module, the load prediction and decision module, and the dynamic power distribution execution module respectively, and is used to provide a multi-level safety response mechanism and ensure reliable communication between modules in the system. By constructing a complete closed-loop control system that includes four major functions: perception, decision-making, execution, and protection, it is used to systematically solve problems such as three-phase imbalance, transformer overload, and harmonic pollution in the power distribution system caused by random start-stop and power fluctuations in charging pile clusters, thereby improving the safety and stability of the power grid and optimizing charging efficiency. The data acquisition module includes a multimodal sensing unit, a synchronous acquisition engine, and a harmonic analysis unit. The multimodal sensing unit is configured at each charging pile power supply node to collect three-phase current, voltage, power factor, and equipment temperature parameters. The synchronous acquisition engine is used to achieve data clock synchronization across power supply nodes. The harmonic analysis unit is used to analyze the harmonic components in the current and voltage waveforms in real time. By setting up the multimodal sensing unit, synchronous acquisition engine, and harmonic analysis unit, a comprehensive data foundation with high accuracy, high synchronization, and power quality details is provided for subsequent load prediction and decision-making. This ensures the accuracy and real-time nature of the information on which the system decisions depend, creating conditions for precise control. The load forecasting and decision-making module includes a hybrid forecasting model and an edge computing controller. The hybrid forecasting model is composed of an LSTM neural network and an ARIMA algorithm fused with dynamic weights to generate load forecast curves for the next 15-30 minutes. The edge computing controller is used to run the hybrid forecasting model and execute the model predictive control algorithm to generate a real-time power distribution control strategy that includes relay timing parameters and capacitor compensation schemes. By adopting the hybrid forecasting model and the edge computing controller, accurate short-term forecasts of charging load trends can be achieved, and the optimal control strategy can be quickly generated at the edge side near the data source based on the forecast results, thereby overcoming the problem of response lag in traditional solutions and realizing forward-looking dynamic power regulation. The dynamic power distribution execution module includes a three-phase balance control unit and a harmonic compensation unit. The three-phase balance control unit is used to control the solid-state relay array to realize the dynamic migration of the charging pile between different phase lines according to the strategy generated by the edge computing controller. The harmonic compensation unit is used to control the switching of the parallel compensation capacitor bank according to the harmonic analysis results. By setting up the three-phase balance control unit and the harmonic compensation unit, the control strategy can be executed directly and quickly, specifically realizing the dynamic switching of the charging pile phase lines to balance the three-phase load, and at the same time compensating for grid harmonics, thereby effectively reducing transformer losses and improving power quality. The safety protection and communication module includes a multi-level safety response unit and a multi-mode communication unit. The multi-level safety response unit is used to implement gradient cooling control, charging priority screening, or self-healing grid reconfiguration according to the severity of the anomaly. The multi-mode communication unit integrates power line carrier communication and wireless communication links for reliable data transmission within the system and with the upper-level platform. By setting up multi-level safety response units and multi-mode communication units, it is ensured that the system can take differentiated protection measures according to the severity of the situation when abnormal operating conditions such as over-temperature, overload, or equipment failure occur. The redundant communication links ensure the reliable transmission of control commands, greatly enhancing the robustness and self-healing capability of the system. The dynamic weight allocation rule in the hybrid prediction model is as follows: when the system detects that the load fluctuation rate is greater than a set threshold, the weight ratio of the LSTM neural network is automatically increased; when the load change is stable, the weight ratio of the ARIMA algorithm is increased. By defining the dynamic weight allocation rule of the hybrid prediction model, the prediction model can adapt to the dynamic change characteristics of the load. When the load changes abruptly, the LSTM neural network is used first to handle nonlinear characteristics, and when the load is stable, the ARIMA algorithm is used first to capture periodic patterns, thereby maintaining high prediction accuracy under various working conditions. The constraints for the three-phase balance control unit to perform phase migration include the source phase line load deviation exceeding 15%, the target phase line remaining capacity being greater than 120% of the power demand of the charging pile to be migrated, and the number of charging piles migrated in a single batch not exceeding 20% ​​of the total number of the cluster. By setting specific constraints for phase migration, the three-phase balance control unit is guided to make decisions on the premise of ensuring the safe and effective migration operation, avoiding new imbalances or equipment overload caused by the migration operation itself, and achieving a stable and controllable load balancing process. The charging priority screening mechanism in the multi-level safety response unit generates an interruption sequence based on the user's historical credit score, the vehicle's charging urgency, and the battery's health status. By clarifying the judgment dimensions for charging priority screening, it can fairly, reasonably, and intelligently determine the order of suspending charging services when the system must reduce its load, prioritizing the emergency charging needs of important users while also taking into account battery health, thereby improving user satisfaction while ensuring system safety. When the multi-mode communication unit detects that the quality of the primary communication channel is lower than the threshold, it can automatically switch to the backup wireless communication link. By specifying the automatic switching function of the communication link, the system can seamlessly switch to the backup wireless link when the quality of the primary communication channel (such as power line carrier) deteriorates due to interference or other reasons, ensuring that key monitoring data and control commands are not interrupted and maintaining the coordinated operation between various modules of the system. It also includes an interface for connecting to a virtual power plant management cloud platform, which is used to receive control instructions from the load aggregation platform and report system operation data. By setting up an interface with the virtual power plant platform, the local charging pile cluster can be used as a flexible and adjustable aggregated resource to participate in a wider range of power grid dispatch and demand response, thereby improving the economy and reliability of the entire power system and expanding the application value of the system.

[0022] The present invention provides the following embodiments: Example 1: Application of charging stations in typical commercial areas Application scenario background: The underground parking lot of a large commercial complex is equipped with 30 7kW AC charging piles, with 10 piles connected to each phase line; during the evening shopping peak, the charging demand is concentrated and the vehicle entry time is random, which can easily lead to a serious imbalance of the three-phase load. System implementation process: Data Acquisition: The multi-modal sensing unit (such as the HTA800 wideband current transformer) in the system's data acquisition module monitors in real time that the current in phase A has risen sharply to 185A, phase B to 150A, and phase C to 140A, and the three-phase imbalance has exceeded 15%; the synchronous acquisition engine ensures that the data timestamps of all nodes are synchronized. Prediction and Decision: The hybrid prediction model (LSTM-ARIMA) in the load prediction and decision module predicts that the load on phase A will remain high for the next 15 minutes based on historical data; the edge computing controller (using an industrial-grade edge computing box of Intel Atom x6425E) then runs the model predictive control algorithm to generate a control strategy: to relocate the two charging piles (pile A05 and A08) on phase A to phase C; Dynamic execution: After receiving the strategy, the three-phase balance control unit of the dynamic power distribution execution module drives the solid-state relay array (such as model CPC1984B) to perform the switching operation; during the switching process, reverse canceling current is injected to suppress inrush current; at the same time, the harmonic compensation unit detects that the 5th harmonic content is slightly high and puts in a set of compensation capacitors. Safety and Communication: Throughout the process, the multimodal communication unit of the safety protection and communication module (using an HPLC+LoRa dual-mode communication module) ensured reliable command transmission; the system monitored that the temperature of each node was normal and no safety response was triggered. Implementation results: Within seconds, the system adjusted the three-phase current to 165A for phase A, 155A for phase B, and 160A for phase C, reducing the imbalance to within 3%; effectively avoiding neutral line overcurrent, reducing transformer losses, and ensuring a normal charging experience for all users.

[0023] Example 2: Handling sudden load surges and equipment failures Application scenario background: At a highway service area, multiple electric buses simultaneously enter the station for fast charging, causing the total load to instantly exceed 95% of the transformer's rated capacity, and a cooling fan failure of one charging pile causes the equipment temperature to rise rapidly to 78°C. System implementation process: Anomaly detection: The temperature sensor of the data acquisition module detected an overheating signal from the faulty charging pile, while the current sensor detected that the total load rate exceeded the limit. Multi-level security response: The multi-level security response unit of the security protection and communication module is triggered; First (gradient cooling control): Gradient cooling is implemented on the phase line where the faulty charging pile is located, gradually reducing its output power from 60kW to 45kW. Secondly (charging priority screening): Due to the still high total load, the system automatically suspended the charging piles serving cars with higher SOC (SOC>70%) based on the charging priority screening mechanism (based on user VIP level and vehicle remaining battery SOC<20% as high urgency). Simultaneously (self-healing grid reconfiguration): the edge computing controller recalculates the power supply topology, dynamically adjusting some loads to the phase lines with lower loads; Communication assurance: During this period, due to the large electromagnetic interference on site, the signal-to-noise ratio of the main HPLC channel of the multimodal communication unit dropped to 12dB. The system automatically and seamlessly switched to the backup LoRa wireless link to ensure the continuous issuance of control commands. Implementation results: The system stabilized the overall load rate within 85% within 1 minute, and the temperature of the faulty charging pile dropped to below 65℃ within 3 minutes, avoiding a potential equipment burnout accident caused by overload and overheating, and prioritizing the charging needs of users who urgently needed charging.

[0024] Example 3: Participation in the Coordinated Scheduling of Virtual Power Plants (VPPs) Application scenario background: This charging pile cluster, as a distributed flexible resource, is connected to the virtual power plant management cloud platform; during the peak electricity consumption period at noon, the VPP platform issues a power limit control instruction of 10% to the charging station; System implementation process: Command reception: The system receives control commands through the interface with the virtual power plant management cloud platform (using the standard IEC 104 communication protocol); Intelligent decision-making and execution: The load prediction and decision-making module does not simply reduce the power of all charging piles uniformly; instead, it combines the prediction results of the hybrid prediction model to determine that some vehicles will be fully charged within the next 30 minutes. The edge computing controller formulates an optimization strategy: maintains the original power for vehicles that are about to be fully charged (SOC>90%), makes only a small power adjustment (such as a 5% reduction) for vehicles that have just been connected (SOC<50%), and focuses on adjusting the power of vehicles in the middle of the charging stage; the dynamic power distribution execution module performs precise control accordingly. Data reporting: During the adjustment process, the system will report real-time running data, including total load, execution results, three-phase balance, etc., to the VPP platform through the multi-modal communication unit; Implementation Results: The charging station responded precisely to the grid's dispatching needs in a way that minimized the impact on users, reducing the total load by 10.5% within the specified time. At the same time, the system maintained its three-phase balance through internal optimization and did not introduce new power quality problems due to responding to grid dispatching. This demonstrates the application value of the system of this invention in a higher-level energy internet.

[0025] Through three systematic real-time embodiments, the effectiveness of the present invention under different operating conditions is systematically verified, fully demonstrating its significant advantages in achieving dynamic load balancing, ensuring safe operation, and participating in grid coordinated dispatch.

[0026] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A dynamic load balancing power supply system for new energy vehicle charging pile clusters, characterized in that: The system includes a data acquisition module, a load prediction and decision-making module, a dynamic power distribution execution module, and a safety protection and communication module. The data acquisition module is used to monitor the three-phase electrical parameters, temperature, and power quality data of the charging pile cluster in real time. The load prediction and decision-making module is connected to the data acquisition module and is used to predict load trends based on historical and real-time data and generate dynamic control strategies. The dynamic power distribution execution module is connected to the load prediction and decision-making module and is used to quickly control power distribution according to the control strategy to achieve three-phase load balancing. The safety protection and communication module is connected to the data acquisition module, the load prediction and decision-making module, and the dynamic power distribution execution module respectively, and is used to provide a multi-level safety response mechanism and ensure reliable communication between the modules in the system.

2. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 1, characterized in that: The data acquisition module includes a multimodal sensing unit, a synchronous acquisition engine, and a harmonic analysis unit. The multimodal sensing unit is configured at each charging pile power supply node to collect three-phase current, voltage, power factor, and equipment temperature parameters. The synchronous acquisition engine is used to achieve data clock synchronization across power supply nodes. The harmonic analysis unit is used to analyze the harmonic components in the current and voltage waveforms in real time.

3. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 2, characterized in that: The load forecasting and decision-making module includes a hybrid forecasting model and an edge computing controller. The hybrid forecasting model is composed of an LSTM neural network and an ARIMA algorithm fused with dynamic weights to generate load forecast curves for the next 15-30 minutes. The edge computing controller is used to run the hybrid forecasting model and execute the model predictive control algorithm to generate a real-time power distribution control strategy that includes relay timing parameters and capacitor compensation schemes.

4. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 3, characterized in that: The dynamic power distribution execution module includes a three-phase balance control unit and a harmonic compensation unit. The three-phase balance control unit is used to control the solid-state relay array to realize the dynamic migration of the charging pile between different phase lines according to the strategy generated by the edge computing controller. The harmonic compensation unit is used to control the switching of the parallel compensation capacitor bank according to the harmonic analysis results.

5. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 4, characterized in that: The safety protection and communication module includes a multi-level safety response unit and a multi-mode communication unit. The multi-level safety response unit is used to implement gradient cooling control, charging priority screening, or self-healing grid reconfiguration according to the severity of the anomaly. The multi-mode communication unit integrates power line carrier communication and wireless communication links for reliable data transmission within the system and between the system and the upper-level platform.

6. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 3, characterized in that: The dynamic weight allocation rule in the hybrid prediction model is as follows: when the system detects that the load volatility is greater than a set threshold, the weight ratio of the LSTM neural network is automatically increased; when the load change is stable, the weight ratio of the ARIMA algorithm is increased.

7. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 4, characterized in that: The constraints for the three-phase balance control unit to perform phase migration include the source phase line load deviation exceeding 15%, the target phase line remaining capacity being greater than 120% of the power demand of the charging pile to be migrated, and the number of charging piles migrated in a single batch not exceeding 20% ​​of the total number of the cluster.

8. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 5, characterized in that: The charging priority screening mechanism in the multi-level safety response unit generates an interruption sequence based on the user's historical credit score, the vehicle's charging urgency, and the battery's health status.

9. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 5, characterized in that: When the multi-mode communication unit detects that the quality of the primary communication channel is below a threshold, it can automatically switch to the backup wireless communication link.

10. The dynamic load balancing power supply system for new energy vehicle charging pile clusters according to claim 1, characterized in that: It also includes an interface for connecting to the virtual power plant management cloud platform, which is used to receive control instructions from the load aggregation platform and report system operation data.

Images