Charging control method and system for electric vehicle in power distribution network

Abstract

The embodiment of the invention discloses a charging control method and system for an electric vehicle in a power distribution network. The system comprises a local control unit and a cloud optimization unit. The method comprises the following steps: a local control unit obtains power distribution network data, user fuzzy charging demand data, a battery state and complex impedance data; the aging state of the battery is analyzed based on the complex impedance data, a real-time health factor is determined, and then the battery loss cost during charging is calculated; and generating a schedulable feature vector representing the schedulable capability of the electric vehicle based on the battery loss cost and various data acquired by the local control unit. The cloud optimization unit determines a category cluster based on the schedulable feature vector; and solving to obtain the cluster scheduling price of each category of clusters by taking minimization of a system objective function as an objective. And the local control unit formulates a charging and discharging plan through a decision-making mechanism according to the cluster scheduling price and the schedulable feature vector, and executes the charging and discharging plan according to the battery state and the communication state with the cloud.

Description

Technical Field

[0001] This application relates to the field of coordinated optimization control of electric vehicles and power distribution networks, and relates to, but is not limited to, a charging control method and system for electric vehicles in a power distribution network. Background Technology

[0002] With the advancement of global carbon peaking and carbon neutrality, the integration of the energy internet and transportation electrification has become an inevitable trend. Electric vehicles, as the core carrier, are evolving from simple transportation tools into distributed mobile energy storage units. The number of electric vehicles is expected to grow exponentially, and the massive, dispersed, and highly random electric vehicle loads are placing unprecedented pressure on the planning and operation of traditional power distribution networks. Traditional power distribution network design relies on load forecasting based on statistical macroscopic laws, resulting in relatively stable load characteristics. However, electric vehicle charging loads exhibit significant spatiotemporal uncertainty, especially in large residential communities where peak residential electricity consumption and peak private car charging times easily overlap during the evening rush hour. Without effective guidance, this double-peak overlap can cause a sudden surge in the load rate of distribution transformers, leading to equipment damage and deterioration of power quality.

[0003] Among related technologies, price-inducing demand response technology guides users to voluntarily charge during off-peak hours through economic levers such as time-of-use pricing. Model-based centralized optimization control technology relies on a cloud computing center to collect precise data such as the real-time battery status of all electric vehicles on the network and user travel plans. By solving a global optimization model, it issues scheduling commands to each charging station. This centralized architecture highly depends on a continuous and stable communication link to maintain the control loop.

[0004] However, the battery aging models relied upon by centralized optimization control technologies are typically static and linear, failing to accurately reflect the nonlinear aging characteristics of lithium-ion batteries during actual operation due to complex electrochemical processes. This can lead to scheduling commands potentially accelerating battery damage or wasting energy storage resources. Cloud computing centers collect sensitive data such as users' precise travel routes, posing a risk of user privacy leaks. As the scale of connected electric vehicles increases, solving high-dimensional optimization problems in the cloud faces computational bottlenecks, making it difficult to meet the real-time scheduling requirements of the distribution network. Because the entire system's control logic heavily relies on instructions from the cloud computing center and lacks local autonomous decision-making capabilities, there is no device-level physical defense mechanism in the event of base station failures or network attacks causing communication interruptions, potentially leading to localized distribution network collapses due to control malfunctions. Existing optimization strategies generally ignore the impact of the physical environment, particularly the balance between charging safety and efficiency in low-temperature environments. Summary of the Invention

[0005] In view of this, embodiments of this application provide a charging control method and system for electric vehicles in a power distribution network, which at least solves the problems of charging control commands causing damage to electric vehicle batteries and cloud computing centers collecting sensitive user data leading to privacy leaks.

[0006] The technical solution of this application embodiment is implemented as follows: In a first aspect, embodiments of this application provide a charging control method for electric vehicles in a power distribution network, applied to a charging control system for electric vehicles in a power distribution network; the system includes: a local control unit and a cloud optimization unit; the method includes: The local control unit is used to acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery status data, and battery complex impedance data. Using the local control unit, based on the battery complex impedance data, the battery aging data of the current electric vehicle is obtained; based on the battery aging data, the real-time health factor of the current battery is determined; based on the real-time health factor, the battery loss cost is calculated through a pre-built battery aging cost model; based on the battery loss cost, the user fuzzy charging demand data, and the current battery state data, the schedulable feature vector of the current electric vehicle is obtained. Using the cloud optimization unit, based on the schedulable feature vector, the category cluster corresponding to the current electric vehicle is determined through preset classification rules; based on the category cluster and the schedulable feature vector, with the optimization objective being to minimize the value of a pre-constructed target optimization function, the target optimization function is solved to obtain the cluster scheduling price corresponding to the category cluster; Using the local control unit, the cluster scheduling price is received, and based on the cluster scheduling price and the schedulable feature vector, the charging and discharging plan of the current electric vehicle is determined through a preset decision-making mechanism; The charging and discharging plan is executed using the local control unit based on the current battery status data and the communication status between the local control unit and the cloud optimization unit.

[0007] Secondly, embodiments of this application provide a charging control system for electric vehicles in a power distribution network, including: a local control unit and a cloud optimization unit; The local control unit is used to acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery status data, and battery complex impedance data; The local control unit is further configured to obtain the battery aging data of the current electric vehicle based on the battery complex impedance data; determine the real-time health factor of the current battery based on the battery aging data; calculate the battery loss cost based on the real-time health factor through a pre-built battery aging cost model; and obtain the schedulable feature vector of the current electric vehicle based on the battery loss cost, the user fuzzy charging demand data, and the current battery state data. The cloud optimization unit is used to determine the category cluster corresponding to the current electric vehicle based on the schedulable feature vector and through preset classification rules; based on the category cluster and the schedulable feature vector, with the optimization objective of minimizing the value of a pre-constructed target optimization function, solve the target optimization function to obtain the cluster scheduling price corresponding to the category cluster; The local control unit is used to receive the cluster scheduling price and, based on the cluster scheduling price and the schedulable feature vector, determine the charging and discharging plan of the current electric vehicle through a preset decision-making mechanism. The local control unit is used to execute the charging and discharging plan based on the current battery status data and the communication status between the local control unit and the cloud optimization unit.

[0008] The beneficial effects of the technical solutions provided in this application include at least the following: This application utilizes a local control unit to acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery state data, and battery complex impedance data, providing a real-time data foundation for formulating charging and discharging plans for current electric vehicles. The complex impedance data reflects the current microscopic health status of the battery. Using the local control unit, based on the battery complex impedance data, the application obtains battery aging data for the current electric vehicle; based on the battery aging data, it determines the real-time health factor of the current battery; based on the real-time health factor, it calculates the battery loss cost using a pre-built battery aging cost model; and based on the battery loss cost, user fuzzy charging demand data, and current battery state data, it obtains the schedulable feature vector of the current electric vehicle. The schedulable feature vector is obtained by fusing user fuzzy charging demand data, current battery state data, and battery complex impedance data through the local control unit. The cloud optimization unit does not directly acquire user fuzzy charging demand data, current battery state data, and battery complex impedance data, avoiding the privacy leakage risk arising from the cloud optimization unit directly acquiring user privacy data. Utilizing a cloud-based optimization unit, based on schedulable feature vectors and pre-defined classification rules, the current electric vehicle's category cluster is determined. Based on the category cluster and schedulable feature vectors, the optimization objective is to minimize a pre-constructed target optimization function, yielding the cluster scheduling price corresponding to each category cluster. By classifying electric vehicles according to pre-defined rules, the optimization problem of electric vehicles in the distribution network is simplified to scheduling category clusters, allowing for differentiated pricing for different clusters and achieving precise and efficient load control of the distribution network. Using a local control unit, the cluster scheduling price is received. Based on the cluster scheduling price and schedulable feature vectors, a pre-defined decision-making mechanism determines the current electric vehicle's charging and discharging plan. This approach respects individual user differences and real-time needs while ensuring the overall goals of the distribution network. The charging and discharging plan is completed locally by the control unit, eliminating the need to upload user privacy data to the cloud-based optimization unit, thus avoiding the risk of data transmission and leakage and protecting user data security. The charging and discharging plan is executed locally by the control unit based on the current battery status data and the communication status between the local control unit and the cloud-based optimization unit. By using the current battery status data as a constraint, potential battery safety risks from executing charge / discharge plans during battery malfunctions are prevented. Continuous monitoring of communication status allows the system to switch to a mode controlled by the local control unit when the local control unit loses connection with the cloud optimization unit. This avoids power grid paralysis caused by cloud optimization unit failures, improves the availability and fault tolerance of the charge / discharge plan, and achieves dynamic intelligent adaptation of the charge / discharge plan. Attached Figure Description

[0009] 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 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, wherein: Figure 1 A schematic flowchart illustrating a charging control method for electric vehicles in a power distribution network, provided as an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a charging control system for an electric vehicle in a power distribution network, provided as an embodiment of this application. Detailed Implementation

[0010] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, 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, not all embodiments. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0011] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0012] It should be noted that the terms "first, second, and third" used in the embodiments of this application are merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, and third" can be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0013] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiments of this application pertain. It should also be understood that terms such as those defined in general dictionaries should be understood to have a meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0014] This application provides a charging control method for electric vehicles in a power distribution network, which is applied to the charging control system of electric vehicles in a power distribution network; the system includes: a local control unit and a cloud optimization unit. Figure 1 A flowchart illustrating a charging control method for electric vehicles in a power distribution network, as provided in this application embodiment, is shown below. Figure 1 As shown, the method includes at least the following steps: Step S110: Using the local control unit, acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery status data, and battery complex impedance data.

[0015] The local control unit can be a smart terminal deployed on the charging side of the current electric vehicle, or it can act as a processor for the smart terminal on the charging side of the current electric vehicle. For example, the smart terminal can be a charging pile.

[0016] Based on distribution network data, the real-time status of the distribution network can be obtained, ensuring that the formulated charging and discharging plans can support the stability of the distribution network. For example, distribution network data can include the voltage of the distribution network.

[0017] Battery complex impedance data is a physical signal that reflects the internal electrochemical state of a battery. Obtaining battery complex impedance data can determine the health status of electric vehicle batteries and avoid increasing battery aging costs due to planned charge and discharge cycles.

[0018] Data from the power distribution network, current user fuzzy charging demand data for electric vehicles, current battery status data, and battery complex impedance data provide a reliable data foundation for subsequent charging and discharging planning.

[0019] Step S120: Using the local control unit, based on the battery complex impedance data, obtain the battery aging data of the current electric vehicle; based on the battery aging data, determine the real-time health factor of the current battery; based on the real-time health factor, calculate the battery loss cost through a pre-built battery aging cost model; based on the battery loss cost, the user fuzzy charging demand data, and the current battery state data, obtain the schedulable feature vector of the current electric vehicle.

[0020] Battery aging data is calculated based on battery complex impedance data and directly reflects the degree of electrochemical aging within the battery. The main component of this data is charge transfer impedance. As the battery ages, the charge transfer impedance gradually increases.

[0021] Real-time health factors can quantify the degree of degradation of a battery's current health status relative to its brand-new state.

[0022] During charging and discharging, batteries experience energy fluctuations, leading to battery aging. Battery degradation cost is the equivalent economic cost of battery life reduction caused by each unit of charge or discharge in the current healthy state. When the real-time health factor indicates the battery is healthy, the cost of charging and discharging degradation is the base cost. When the real-time health factor indicates the battery is aging or in a poor condition, the cost of degradation may be proportionally amplified.

[0023] By calculating the aging cost of the current battery in real time using complex impedance data, the scheduling system can obtain the impact of each charge and discharge cycle on the current battery life. This avoids charging and discharging plans that are too conservative, leading to low charging efficiency, or too aggressive, accelerating battery health degradation.

[0024] The local control unit integrates user fuzzy charging demand data, current battery status data, and battery complex impedance data to obtain a schedulable feature vector. The cloud-based optimization unit does not directly access these data, thus avoiding the privacy risks associated with directly accessing user data.

[0025] Step S130: Using the cloud optimization unit, based on the schedulable feature vector and through preset classification rules, determine the category cluster corresponding to the current electric vehicle; based on the category cluster and the schedulable feature vector, with the optimization objective being to minimize the value of a pre-constructed target optimization function, solve the target optimization function to obtain the cluster scheduling price corresponding to the category cluster.

[0026] The cloud-based optimization unit receives the schedulable feature vectors of all electric vehicles within its control area. Based on the similarity of the schedulable feature vectors corresponding to each electric vehicle, the cloud-based optimization unit classifies all electric vehicles into different category clusters using preset classification rules. For example, the preset classification rules could be K-Means clustering or Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm. For instance, all vehicles that are not in a hurry to charge and have healthy batteries are grouped into a high-flexibility cluster, while all vehicles that urgently need charging are grouped into a high-urgency cluster.

[0027] The cloud-based optimization unit operates on a cluster basis, representing all electric vehicles within the control area and corresponding to different categories. Based on the schedulable feature vector of each cluster and the global objective of the distribution network, it runs an optimization algorithm to solve for the objective function that minimizes the total system cost. For example, the schedulable feature vector could represent a high-flexibility cluster, which has a larger average schedulable power; the optimization algorithm could be a Pythagorean optimization algorithm.

[0028] The objective optimization function outputs different cluster scheduling prices for each cluster category. For example, for high-flexibility clusters, a high electricity price is output to encourage them to discharge, while a low electricity price is output to high-urgency clusters to ensure they are charged.

[0029] For highly flexible clusters, when the battery state of electric vehicles is high and users have ample charging time, the objective optimization function will generate a higher cluster scheduling price. This cluster scheduling price makes the urgency index of users within the cluster lower than the cluster scheduling price, thereby triggering the economic-driven strategy on the edge side. This guides such clusters to suppress charging or actively feed power to the grid during peak grid load periods, providing key peak-shaving support for the grid and creating additional economic benefits for users.

[0030] For high-urgency clusters, when electric vehicles within the cluster generally have low state of charge or users' set off-grid times are approaching, triggering urgent charging needs, the objective optimization function generates a cluster scheduling price lower than the cluster's average urgency index. This price signal ensures that the edge side meets the rigid condition of "urgency greater than price," thereby triggering the edge controller's greedy execution strategy. This strategy allows the controller to ignore current economic costs, ensuring that electric vehicles can quickly replenish their power with the highest power priority allowed by the equipment, thus prioritizing users' urgent travel needs.

[0031] In situations where the distribution network is experiencing low load, such as when the system detects a decrease in total charging demand and a significant margin in the distribution network's power supply capacity, the cloud-based optimization unit lowers the cluster dispatch price for each type of cluster. This price reduction lowers the rigidity threshold for determining clusters, incentivizing all types of clusters to increase charging power for energy storage. This is especially true for price-sensitive clusters with greater dispatchability, which increase charging power for energy storage. This enhances the distribution network's ability to absorb intermittent renewable energy sources such as wind and solar power, thereby improving the overall economic efficiency of the power system.

[0032] By classifying electric vehicles according to preset classification rules, the corresponding category clusters of the current electric vehicles are obtained. This simplifies the optimization problem of electric vehicles in the distribution network to the scheduling of category clusters, thus overcoming the bottleneck of centralized computing power. Differentiated prices are set for different category clusters, enabling precise and efficient control of the distribution network load.

[0033] Step S140: Using the local control unit, receive the cluster scheduling price, and based on the cluster scheduling price and the schedulable feature vector, determine the charging and discharging plan of the current electric vehicle through a preset decision-making mechanism.

[0034] For example, even if the cluster scheduling price output by the cloud optimization unit is high, indicating encouragement to discharge, the updated schedulable feature vector shows that the user's need for the vehicle is high. The local control unit will change the current charging and discharging plan of the electric vehicle to one of the following: encouraging discharge, discharging only a small amount, or charging, in order to prioritize the user's needs.

[0035] While ensuring the overall goals of the power distribution network, it respects individual user differences and real-time needs to the greatest extent possible. Furthermore, current electric vehicle charging and discharging plans are completed based on local control units using local data, eliminating the need to continuously upload user privacy data to cloud-based optimization units. This reduces the exposure and transmission of user privacy data, protecting user data security.

[0036] Step S150: Using the local control unit, the charging and discharging plan is executed based on the current battery status data and the communication status between the local control unit and the cloud optimization unit.

[0037] When executing a charge / discharge plan, the local control unit also needs to dynamically adjust the plan based on the current battery status data and the communication status between the local control unit and the cloud optimization unit. The current battery status data monitors the internal environment of the electric vehicle. This data is dynamically changing, and the execution of the charge / discharge plan must ensure the safety of the battery. The communication status monitors the control commands from the cloud optimization unit, reflecting whether these commands are reliably and continuously delivered.

[0038] Using the current battery status data as a constraint prevents potential battery safety risks from mechanically executing the charge / discharge plan when the battery is abnormal. By continuously monitoring the communication link status, the system switches to a degraded control mode when the local control unit loses connection with the cloud optimization unit. This avoids power grid paralysis caused by a single point of failure in the cloud optimization unit, improves the overall availability and fault tolerance of the charge / discharge plan, and achieves dynamic intelligent adaptation of the charge / discharge plan.

[0039] This application utilizes a local control unit to acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery state data, and battery complex impedance data, providing a real-time data foundation for formulating charging and discharging plans for current electric vehicles. The complex impedance data reflects the current microscopic health status of the battery. Using the local control unit, based on the battery complex impedance data, the application obtains battery aging data for the current electric vehicle; based on the battery aging data, it determines the real-time health factor of the current battery; based on the real-time health factor, it calculates the battery loss cost using a pre-built battery aging cost model; and based on the battery loss cost, user fuzzy charging demand data, and current battery state data, it obtains the schedulable feature vector of the current electric vehicle. The schedulable feature vector is obtained by fusing user fuzzy charging demand data, current battery state data, and battery complex impedance data through the local control unit. The cloud optimization unit does not directly acquire user fuzzy charging demand data, current battery state data, and battery complex impedance data, avoiding the privacy leakage risk arising from the cloud optimization unit directly acquiring user privacy data. Utilizing a cloud-based optimization unit, based on schedulable feature vectors and pre-defined classification rules, the current electric vehicle's category cluster is determined. Based on the category cluster and schedulable feature vectors, the optimization objective is to minimize a pre-constructed target optimization function, yielding the cluster scheduling price corresponding to each category cluster. By classifying electric vehicles according to pre-defined rules, the optimization problem of electric vehicles in the distribution network is simplified to scheduling category clusters, allowing for differentiated pricing for different clusters and achieving precise and efficient load control of the distribution network. Using a local control unit, the cluster scheduling price is received. Based on the cluster scheduling price and schedulable feature vectors, a pre-defined decision-making mechanism determines the current electric vehicle's charging and discharging plan. This approach respects individual user differences and real-time needs while ensuring the overall goals of the distribution network. The charging and discharging plan is completed locally by the control unit, eliminating the need to upload user privacy data to the cloud-based optimization unit, thus avoiding the risk of data transmission and leakage and protecting user data security. The charging and discharging plan is executed locally by the control unit based on the current battery status data and the communication status between the local control unit and the cloud-based optimization unit. By using the current battery status data as a constraint, potential battery safety risks from executing charge / discharge plans during battery malfunctions are prevented. Continuous monitoring of communication status allows the system to switch to a mode controlled by the local control unit when the local control unit loses connection with the cloud optimization unit. This avoids power grid paralysis caused by cloud optimization unit failures, improves the availability and fault tolerance of the charge / discharge plan, and achieves dynamic intelligent adaptation of the charge / discharge plan.

[0040] Optionally, the method further includes: using the local control unit, when the current battery establishes a charging connection with the local control unit, multiplexing the DC-DC converter built into the local control unit as a signal generator, and applying a sinusoidal AC signal containing multiple different frequency components superimposed on the DC charging current by micro-perturbation modulation of the pulse width modulation (PWM) signal of the power switching transistor in the DC-DC converter and applying it to the current battery; using the local control unit, measuring the response data of the current battery to each of the frequency components to obtain the corresponding voltage response signal and current response signal; using the local control unit, performing denoising processing on the voltage response signal and the current response signal based on the discrete wavelet transform algorithm to obtain denoised voltage response signal and denoised current response signal; using the local control unit, calculating the complex impedance value corresponding to each of the frequency components based on the denoised voltage response signal and the denoised current response signal to obtain the battery complex impedance data.

[0041] To acquire the internal state of the battery in real time during the charging process of an electric vehicle, while avoiding the need for an expensive specialized electrochemical workstation, this application proposes utilizing existing power electronics within the local control unit as a detection device. Specifically, the DC-DC converter within the electric vehicle's power supply equipment is used as a signal generator. By performing micro-perturbation modulation on the pulse width modulation (PWM) signal of the switching transistors in the circuit, a sinusoidal AC signal with a small amplitude and multiple frequency components can be superimposed on the DC current used for charging, resulting in the corresponding voltage and current response signals.

[0042] In some specific implementations, conventional methods typically require an expensive external electrochemical workstation to obtain the battery impedance spectrum. This application proposes a hardware-in-the-loop electrochemical impedance spectroscopy (EIS) injection method. Specifically, the system directly utilizes the DC-DC converter within the local control unit (i.e., the charging pile) used to control the flow of electrical energy, creatively reusing it as a signal generator. During charging, the microcontroller applies a high-frequency micro-perturbation to the duty cycle of the pulse width modulation (PWM) signal driving the power switching transistors (such as IGBTs or MOSFETs). This minute perturbation at the control signal level, under the action of the filter inductor on the converter output side, can be smoothly applied to the main charging DC power at the hundred-ampere level. On top of this, multiple sinusoidal AC components with extremely small amplitudes (not causing polarization damage to the battery lattice structure) and covering a wide frequency range from 0.1 Hz to 1 kHz are superimposed. This physical layer design based on topology reuse completely eliminates the dependence on external electrochemical detection hardware, realizing millisecond-level microscopic physical state sensing with zero hardware increase cost.

[0043] The sinusoidal AC signal is shown in formula (1): Formula (1); in, Represents a sinusoidal alternating current signal. Represents the current time. This represents the DC bias current component. The value represents the amplitude of the sinusoidal alternating current component at the m-th frequency. This represents the angular frequency of the m-th sinusoidal AC component. This represents the initial phase of the m-th frequency sinusoidal AC component.

[0044] frequency The scanning range is Hertz (unit: Hz) to Hertz. The impedance spectrum within this scanning range can reflect the current health status of the battery.

[0045] In some embodiments, applying a sinusoidal AC signal containing multiple different frequency components to the current battery includes applying a single frequency signal sequentially, or applying a composite signal consisting of multiple superimposed frequencies simultaneously.

[0046] However, the internal resistance of the battery and the resistance of the measurement circuit can interfere with the weak AC response signal, affecting the measurement accuracy. During signal acquisition, a high-precision analog front-end chip is used, and the voltage and current response signals at both ends of the battery are simultaneously acquired using a four-wire method. After Fourier transform, the acquired voltage and current response signals yield the battery complex impedance data, as shown in formula (2): Formula (2); in, This represents the battery at an angular frequency of Battery complex impedance data at that time, Represents AC voltage response signal In frequency Complex spectrum value at that location, Represents alternating current excitation signal In frequency Complex spectrum value at that location, represent The real part, represent The imaginary part, It represents the imaginary part of a complex number.

[0047] Optionally, determining the real-time health factor of the current battery based on the battery aging data includes: using the local control unit to obtain the standard value of the charge transfer impedance and the optimal operating temperature of the current battery; using the local control unit to obtain the battery aging data of the current battery by fitting a pre-built equivalent circuit model based on the battery complex impedance data; the battery aging data includes the current charge transfer impedance and the current temperature; using the local control unit to calculate the impedance aging term that deviates from the standard value of the charge transfer impedance; using the local control unit to calculate the temperature offset term that deviates from the optimal operating temperature of the battery; and using the local control unit to fuse the impedance aging term and the temperature offset term through a nonlinear amplification mechanism to obtain the real-time health factor that generates an exponential penalty response when the impedance abnormally increases or the temperature deviates drastically.

[0048] Based on measured battery complex impedance data, the local control unit identifies and obtains battery aging data by fitting a pre-built equivalent circuit model. The battery aging data may include charge transfer impedance.

[0049] Charge transfer impedance is a key indicator characterizing the electrochemical reaction kinetics at the electrode interface. Changes in charge transfer impedance are directly related to the thickening of the solid electrolyte interfacial film and the risk of lithium plating, and are the core basis for assessing the health status of the battery.

[0050] Based on the charge transfer impedance extracted from the battery complex impedance data, the local control unit constructs an adaptively evolving battery aging cost model. By introducing a real-time health factor, the calculation of battery aging cost is dynamically corrected. The real-time health factor consists of an impedance aging term reflecting changes in charge transfer impedance, and a term reflecting the effects of temperature stress.

[0051] The nonlinear amplification mechanism uses an exponential function to exponentially amplify the negative impact of temperature on battery health when calculating the real-time battery health factor, ensuring that the algorithm makes an extremely sensitive and enhanced protective response to extreme temperature conditions.

[0052] Real-time health factors are calculated based on a nonlinear amplification mechanism, as shown in formula (3): Formula (3); in, The value of is greater than or equal to 1. For charge transfer impedance, For battery temperature, This is the impedance aging weighting factor. This is the temperature stress weighting coefficient. To measure the charge transfer impedance, The initial charge transfer impedance, For the optimal operating temperature of the battery, This is the temperature sensitivity coefficient. The optimal operating temperature for a battery is typically 25 degrees Celsius.

[0053] In some embodiments, the operation mechanism of the battery aging cost model can be reflected by the real-time numerical change of the health factor. When the battery is in a healthy state, its real-time health factor approaches the standard value of 1. At this time, the battery loss cost calculated by the battery aging cost model is basically consistent with the calculation results of the traditional static model, thereby ensuring that the local control unit's assessment of the loss of batteries in good condition is accurate and meets expectations. When the battery is in a sub-healthy state, for example, due to a significant increase in its charge transfer impedance or extremely low operating temperature, its real-time health factor is much greater than 1. In this state, the battery aging cost model outputs a sharply increased dynamic battery loss cost. This allows the local control unit to automatically assign a higher virtual cost to batteries in a vulnerable state when formulating charge and discharge plans. Thus, in the optimized scheduling, such batteries are prioritized for protection, avoiding the application of charge and discharge power that may accelerate aging, achieving adaptive protection based on the real-time microscopic health status of the battery.

[0054] The battery aging cost model is shown in formula (4): Formula (4); in, Cost of battery wear and tear during charging. For real-time health factors, For the purchase cost of batteries, Total energy throughput over the battery's lifetime. Depth of discharge The cycle life is below.

[0055] When the charge transfer impedance inside a battery increases abnormally or its operating temperature is at an extremely low level, the battery loss cost during charging will increase dramatically. For example, to achieve the lowest total system cost or the smoothest load curve, high-intensity charge and discharge operations on the current battery are avoided, thus realizing asset protection based on the real-time physical state of the battery.

[0056] Optionally, the method further includes: constructing an initial battery aging model framework; obtaining historical battery aging data and historical health factors; inputting the historical battery aging data and historical health factors into the initial battery aging model framework to obtain battery aging parameters; and determining the pre-constructed battery aging cost model based on the battery aging parameters.

[0057] Based on this pre-built battery aging cost model, economic levers are introduced. The microscopic physical aging of the battery is mapped to a quantifiable battery loss cost in the macroscopic scheduling algorithm through the model. This battery loss cost is integrated into the objective function of the cloud optimization unit and the decision logic of the local control unit, thereby automatically degrading the use of sub-healthy batteries, that is, reducing their charging and discharging priority and power in the scheduling.

[0058] Therefore, by using a physical information coupling mechanism across time scales, the problem of battery over-protection or under-protection caused by static models is solved, thereby improving the utilization efficiency and operational safety of battery assets.

[0059] Optionally, the schedulable feature vector includes an urgency index, a flexibility index, a health confidence level, and a capacity contribution value. Obtaining the schedulable feature vector of the current electric vehicle based on the battery loss cost, the user's fuzzy charging demand data, and the battery's current state data includes: acquiring the rated power corresponding to the local control unit; obtaining the urgency index and the flexibility index based on the user's fuzzy charging demand data and the battery's current state data, respectively; determining the health confidence level representing the battery's willingness to participate in discharge based on the battery loss cost; obtaining the capacity contribution value using the local control unit based on the rated power and the battery's current state data; and aggregating and generating the schedulable feature vector of the current electric vehicle using the local control unit based on the urgency index, the flexibility index, the health confidence level, and the capacity contribution value.

[0060] In some embodiments, the schedulable feature vector is a nonlinear feature map used to hide the user's actual travel data, which includes an urgency index, a flexibility index, a health confidence level, and a capacity contribution value. Based on the user's fuzzy charging demand data and the battery's current state data, a nonlinear data anonymization calculation is performed to obtain the urgency index, which characterizes the rigidity of the user's demand but does not include a specific time, and the flexibility index, which characterizes the acceptable power reduction range.

[0061] Battery current status data can include the battery's current state of charge and maximum charging power.

[0062] The local control unit acquires the user's fuzzy charging demand data and the current battery status data, and calculates the schedulable feature vector. The calculation is shown in formula (5): Formula (5); in, This is an urgency index, a normalized scalar, representing the rigidity of user needs. As a flexibility index, For health confidence, Contribution to capacity.

[0063] The flexibility index characterizes the acceptable range of charging power adjustments or charging time shifting windows for a current electric vehicle. The local control unit obtains real-time battery status data from the vehicle's battery management system, including the current state of charge and the maximum allowable charging power. Based on the current battery status data and the user's fuzzy electricity demand data, the flexibility index is obtained by calculating the minimum demand time and the available time window.

[0064] Health confidence reflects the degree to which an electric vehicle is willing to participate in discharge regulation. It is calculated based on battery degradation costs derived from a battery aging cost model. Higher battery degradation costs correspond to lower health confidence values, indicating a lower willingness for the battery to participate in discharge under its current condition.

[0065] The capacity contribution value characterizes the maximum instantaneous capacity boundary that an electric vehicle can participate in power regulation of the distribution network at the current moment. The capacity contribution value equals the product of the charging pile's rated maximum power and the battery's available state of charge (SOC). The rated power of the charging pile corresponding to the local control unit is a fixed attribute of the charging pile hardware, while the SOC margin is calculated based on real-time data provided by the battery management system, specifically the difference between the set upper and lower SOC limits. The capacity contribution value reflects the hardware's power capability and the battery's real-time energy reserves. This application performs nonlinear characteristic transformation at the data generation source, cutting off the path to user privacy leakage at the system architecture level. Simultaneously, it reduces the real-time requirements for communication bandwidth, promoting the deployment of large-scale vehicle-to-grid collaborative control applications in narrowband IoT and other communication networks.

[0066] The mechanism for extracting and generating schedulable feature vectors fundamentally reconstructs the privacy protection paradigm at the system's underlying architecture. Unlike existing technologies that employ symmetric encrypted pass-through (such as encrypted transmission of original departure time and target node information), this application performs an irreversible nonlinear feature transformation at the source of data generation (the local control unit). For example, it maps the highly specific expected off-grid time and target state of charge to a normalized (between 0 and 1) scalar "urgency index." Because this mapping process loses the specific time and spatial dimensions, the cloud optimization unit receives only a one-dimensional abstract rigidity index, completely blocking the possibility of reverse engineering to deduce the user's real-life trajectory or address information from a mathematical information theory perspective. Furthermore, by transmitting only low-frequency schedulable feature vectors to the cloud, rather than high-frequency telemetry of massive real-time power and voltage operating curves, the data communication bandwidth requirements of this system decrease exponentially. Subsequently, the cloud further utilized pre-defined classification rules such as density-based noisy application space clustering (DBSCAN) to map massive discrete vehicle nodes into a few category clusters for joint shadow pricing optimization. This successfully reduced the original problem, which had millions of variables and faced the risk of combinatorial explosion, to a constant-level optimization space, thus overcoming the computing power bottleneck and dimensionality curse problem of ultra-large-scale centralized scheduling.

[0067] Optionally, the step of using the cloud optimization unit to determine the category cluster corresponding to the current electric vehicle based on the schedulable feature vector and through preset classification rules; and solving the target optimization function based on the category cluster and the schedulable feature vector, with minimizing the value of a pre-constructed target optimization function as the optimization objective, to obtain the cluster scheduling price corresponding to the category cluster, includes: using the cloud optimization unit to calculate the spatial distance between the schedulable feature vector and the center vector of each preset category cluster; using the cloud optimization unit to determine the category cluster corresponding to the current electric vehicle based on the spatial distance; using the cloud optimization unit to determine the average urgency corresponding to the category cluster based on the schedulable feature vector; using the cloud optimization unit to obtain the real-time cost price of the distribution network; and using the cloud optimization unit to input the real-time cost price and the average urgency into the pre-constructed target optimization function, and solving the target optimization function based on the Horse herd optimization algorithm, with minimizing the value of the pre-constructed target optimization function as the optimization objective, to obtain the cluster scheduling price corresponding to the category cluster and the power purchase of the distribution network.

[0068] The objective optimization function is shown in equation (6): Formula (6); in, For electricity purchase costs, This is a penalty item. For the total cost of the distribution network, For the current time, For the power purchase capacity, Let be the real-time electricity price at time t; k is the cluster index. Let be the total number of clusters, at time t, The shadow price to be sent to the k-th cluster, Let be the average urgency of the k-th cluster.

[0069] After receiving the cluster dispatch price from the cloud optimization unit, the local control unit will make autonomous decisions based on its local schedulable feature vector. This mechanism ensures optimal local response within the cloud-based macro-optimization framework. When the urgency index calculated by the local control unit is higher than the received cluster dispatch price, it indicates that the user's immediate charging demand exceeds the cluster dispatch price offered by the distribution network during that period. In this case, priority is given to meeting the user's core needs, ignoring the cluster dispatch price, and charging the battery at the maximum permissible power.

[0070] When the locally calculated urgency index is lower than or equal to the received cluster dispatch price, it indicates that user demand is elastic and the distribution network has a clear willingness to adjust during this period. Therefore, the local control unit will proactively respond to this price by reducing charging power. If the price is significantly higher than the locally assessed battery loss cost, it will execute a vehicle discharge operation to the distribution network, thereby responding to the distribution network's demand and obtaining corresponding economic incentives.

[0071] Optionally, the schedulable feature vector includes an urgency index; the charging and discharging plan includes a first charging and discharging plan and a second charging and discharging plan; the step of using the local control unit to receive the cluster scheduling price and, based on the cluster scheduling price and the schedulable feature vector, determining the charging and discharging plan of the current electric vehicle through a preset decision mechanism includes: using the local control unit to compare the cluster scheduling price with the urgency index and the battery loss cost when the cluster scheduling price is received; using the local control unit to determine that when the urgency index is greater than the cluster scheduling price, it is determined that rigid demand is dominant, and the first charging and discharging plan is determined to ignore economic costs and charge at the maximum allowable power; using the local control unit to determine that when the urgency index is less than or equal to the cluster scheduling price, it is determined that economy is dominant, and the second charging and discharging plan is determined to respond to the adjustment signal to reduce the charging power; the charging priority of the second charging and discharging plan is higher than that of the first charging and discharging plan.

[0072] In a specific implementation, after receiving the cluster dispatch price from the cloud, the local control unit does not compare it with a fixed price threshold. Instead, it executes a greedy decision-making logic based on multi-dimensional dynamic variable game theory. Specifically: when the locally calculated urgency index is greater than the received cluster dispatch price, the system determines that it is currently in a state dominated by rigid demand. At this time, the first charging and discharging plan is triggered, meaning the system will prioritize the user's travel time value, ignoring the current high grid dispatch price, and forcibly charge the device at its maximum allowed power. Conversely, when the urgency index is less than or equal to the cluster dispatch price, the system determines that the user has time flexibility and enters an economic-driven state, triggering the second charging and discharging plan, actively reducing the charging power to respond to grid peak shaving. Furthermore, within the execution framework of the second charging and discharging plan, the system not only performs passive power suppression but also compares the cluster dispatch price with the battery loss cost dynamically calculated locally based on EIS in real time. Only when the cluster dispatch price (i.e., the economic return of participating in grid regulation) is strictly greater than the battery loss cost converted from a single charging and discharging cycle will the vehicle be driven to perform reverse discharge to the distribution network (V2G). This stringent secondary interception mechanism completely eliminates the underlying economic paradox in existing V2G projects where discharge revenue cannot cover the accelerated degradation of battery life. It achieves true break-even arbitrage at the edge, greatly enhancing user asset security and their enthusiasm for participating in distribution network dispatch.

[0073] After receiving the cluster scheduling price from the cloud optimization unit, the local control unit will perform a greedy decision based on its local schedulable feature vector. This mechanism ensures the optimal response for local units within the cloud-based macro-optimization framework. The specific decision logic is as follows: When the locally calculated urgency index is higher than the received cluster dispatch price, it indicates that the user's charging demand is extremely urgent, and its time value exceeds the current distribution network price signal. In this case, the system will prioritize the user's travel needs, ignoring economic cost factors, and charge the battery at the maximum permissible power.

[0074] When the locally calculated urgency index is lower than or equal to the received cluster scheduling price, it indicates that the user has ample time and scheduling flexibility. In this case, the local control unit will proactively respond to the adjustment signal from the cloud optimization unit and reduce charging power. If the real-time electricity price is higher than the locally assessed battery loss cost, the vehicle will discharge to the power distribution network, participate in power distribution network regulation, and obtain economic benefits.

[0075] Optionally, the current battery status data includes the battery temperature; the step of using the local control unit to execute the charge / discharge plan based on the current battery status data and the communication status between the local control unit and the cloud optimization unit includes: using the local control unit to allocate the heating power and charging power of the current battery based on a preset battery thermal model when the battery temperature is lower than a preset temperature threshold, so as to preheat the current battery; and using the local control unit to execute the charge / discharge plan based on the communication status between the local control unit and the cloud optimization unit after the preheating of the current battery is completed.

[0076] In some embodiments, the current battery state data includes the battery temperature; using a local control unit, based on the current battery state data and the communication status between the local control unit and the cloud optimization unit, a charge / discharge plan is executed, including: using the local control unit, when the battery temperature is lower than a preset temperature threshold, activating an electrothermal multiphysics coupling optimization strategy; based on a preset battery thermal model and the maximum allowable charging power boundary condition that dynamically changes with battery temperature, embedding the thermodynamic dimension into the objective optimization function, and performing time-series collaborative optimization and dynamic allocation of the current battery's heating power and charging power; wherein, collaborative optimization includes a pre-emptive heat storage mode based on electricity price prediction and a Pareto optimal charging mode that balances heating time and charging time based on the user urgency index; using the local control unit, based on the dynamically allocated power command and the communication status with the cloud optimization unit, collaboratively executing the dynamically adjusted heating and charge / discharge plan.

[0077] The current battery status data includes battery temperature and maximum charging power.

[0078] For extreme low-temperature environments, such as when the ambient temperature is below -10 degrees Celsius, the edge control unit initiates an electrothermal coupling optimization strategy. In this scenario, the objective optimization function introduces a battery thermal management energy consumption variable into the energy balance equation. For vehicles with high urgency, power is prioritized to activate the heating system to unlock fast charging capabilities. For vehicles with lower urgency, low-power heat preservation and preheating are performed during off-peak hours. This allows for finding a Pareto optimal solution between heating energy consumption and charging efficiency, thus avoiding damage to battery life caused by low-temperature lithium plating. For extreme low-temperature environments, this application introduces a thermodynamic dimension into the optimization model to establish an electrothermal coupling energy consumption model. The total energy consumption for replenishing energy for vehicles in cold environments is calculated as shown in formula (7): Formula (7); in, Total energy consumption, For charging energy consumption, For heating energy consumption, For ambient temperature, For the target temperature, This represents the available time.

[0079] The differential equation for battery thermal equilibrium is shown in equation (8): Formula (8); in, This refers to the specific heat capacity of the battery. For battery quality, For battery temperature, For time, Represents the rate of temperature change. Represents the power of the external heater. For charging current, Internal resistance The convective heat transfer coefficient between the battery and the air. This is the effective heat dissipation area of ​​the battery. For ambient temperature, For thermal power, This refers to the convective heat dissipation power.

[0080] The maximum allowable charging power as a function of battery temperature is calculated as shown in formula (9): Formula (9); in, The maximum charging power at time t. For the preset mapping function, Let t be the battery temperature at time t.

[0081] Based on the thermodynamic and electrochemical characteristics of batteries, the maximum allowable charging power of a battery decreases significantly with decreasing temperature. The core physical mechanism of this limitation lies in preventing lithium ions from depositing on the negative electrode surface at low temperatures, i.e., lithium plating, which irreversibly damages battery health and safety. Therefore, directly using high-power charging under low-temperature conditions is strictly limited and poses safety risks. To address this issue, this application provides two electrothermal coupling optimization strategies. The first is an off-peak heat storage mode, with an optimization objective prioritizing economy. This strategy utilizes off-peak electricity prices at night to initiate low-power battery heating in advance. The local control unit, based on the user's preset departure time and target temperature, reverse-calculates the optimal heating start time, ensuring that the battery reaches its optimal operating temperature when the user departs, while minimizing total electricity costs. It utilizes the spatiotemporal energy price difference in the distribution network, treating thermal energy as a virtual energy form that can be stored in advance. The second is a Pareto rapid charging mode, designed specifically for users with high urgency. This mode uses Pareto front analysis to compare the total time consumption of two paths: heating before fast charging and slow charging throughout at low temperatures. Because the maximum charging power is extremely low at low temperatures, a moderate heating process is first performed to raise the battery temperature, followed by efficient fast charging. This method typically takes significantly less time than slow charging at low temperatures throughout the entire process. The algorithm automatically compares and selects the charging path with the lowest time cost.

[0082] In extremely cold natural environments (e.g., -10°C and below), the intercalation kinetics of lithium ions on the graphite anode surface are severely suppressed. Blindly applying a large current for charging at this time will lead to irreversible deposition of metallic lithium on the anode surface (i.e., lithium plating). This not only causes permanent capacity collapse damage to the battery, but also, if the separator is punctured, can trigger a serious thermal runaway safety accident. Therefore, the maximum permissible charging power boundary defined by formula (9) is crucial. This is not a static constant, but a strict physical constraint that exhibits a highly nonlinear positive correlation with the real-time battery temperature. The electrothermal multiphysics coupling optimization strategy proposed in this application abandons the crude and simplistic "full-power heating and restart charging" serial logic. During the execution of the Pareto rapid charging mode, the optimization solver built into the local control unit performs multi-objective continuous collaborative optimization in the two-dimensional parameter decision space (i.e., the heating power allocated to the positive temperature coefficient heating circuit (PTC heating loop) in real time and the effective charging power allocated to the battery body). Based on thermodynamic equations, the algorithm dynamically predicts that if more allocated power is injected into the heating system in the initial stage to achieve rapid temperature rise, the charge transfer impedance inside the battery will decrease exponentially, leading to… The charging safety boundary is rapidly and significantly widened outward, enabling it to safely handle high-rate fast charging currents several times greater than the initial state in the latter half of the charging process. By traversing and calculating the total system energy consumption and total dwell time corresponding to massive power combination sequences across multiple time scales, the system autonomously plots and tracks the Pareto Frontier curve, dynamically issuing coupled power allocation commands along the lower edge path with the optimal time cost, thus breaking through the charging physical efficiency bottleneck in extremely cold environments.

[0083] The thermodynamic equations describing battery temperature changes are embedded into the power dispatch model, making the maximum charging power a dynamically changing boundary condition. The local control unit collaboratively optimizes the process, simultaneously deciding on the timing allocation of heating and charging power to achieve both economical heat storage and rapid charging.

[0084] This application breaks through the limitations of traditional vehicle scheduling that only considers power balance. For cold regions, it proposes a method to coordinate virtual heat storage and electric heat power scheduling using battery thermal inertia, which improves the practicality of electric vehicle charging control system in extreme environments.

[0085] Optionally, executing the charging and discharging plan based on the communication status between the local control unit and the cloud optimization unit includes: when communication is normal, executing the charging and discharging plan using the local control unit; when communication is interrupted, determining an updated charging and discharging plan based on the power distribution network data using the local control unit; the updated charging and discharging plan includes reducing the charging power corresponding to the charging and discharging plan and keeping the charging power corresponding to the charging and discharging plan unchanged; when communication is restored, the local control unit connects to the cloud optimization unit after meeting preset communication quality conditions.

[0086] In some embodiments, a charging and discharging plan is executed based on the communication status between the local control unit and the cloud optimization unit, including: when the local control unit continuously receives a cloud signal within a preset heartbeat cycle and determines that the communication is normal, the local control unit is in a collective intelligence mode and executes the charging and discharging plan; when the heartbeat timeout determines that the communication is interrupted, the local control unit switches to an island defense mode, cuts off the dependence on the cloud optimization unit, and activates nonlinear voltage-active power droop control based on local distribution network voltage feedback to determine the updated charging and discharging plan: using a local voltage transformer to collect the grid connection point voltage, when a voltage drop occurs due to overload, the actual charging power is forcibly reduced according to a preset nonlinear exponential law to curb voltage collapse; when communication is restored, the local control unit triggers a chaotic reconnection soft-start mechanism: using Logistic chaotic mapping to iteratively generate non-repeating chaotic sequence values, and mapping the chaotic sequence values ​​to a random delay time for the local control unit to reconnect to the cloud optimization unit, so that the reconnection times of multiple local control units in the distribution network are evenly distributed on the time axis.

[0087] To ensure the system's survivability under extreme failures such as communication interruption, this application designs a collective intelligence mode and an island defense mode, and the local control unit can switch control between the two states of normal communication and communication interruption.

[0088] The Crowd Intelligence Mode is the default state for normal operation. The local control unit enters Crowd Intelligence Mode when it continuously receives valid signals from the cloud-based optimization unit within a preset heartbeat cycle. In Crowd Intelligence Mode, the local control unit executes a charging / discharging plan based on schedulable feature vectors and cluster scheduling prices.

[0089] Islanding defense mode is an emergency fallback state in case of communication interruption. When the local control unit detects that it has failed to receive a communication signal from the cloud optimization unit within a predetermined heartbeat cycle, it determines that a communication failure has occurred. At this time, the local control unit switches to islanding defense mode, cuts off its dependence on the cloud optimization unit, and activates islanding defense mode. Islanding defense mode utilizes distribution network data collected in real time by voltage transformers to execute a nonlinear voltage active power droop control algorithm.

[0090] The calculation of the actual charging power generated by the local control unit is shown in formula (10): Formula (10); in, Represents the actual charging power. For planned power, Where is the rated power, and sgn(ΔV) is the power adjustment function. The voltage deviation between the measured voltage and the rated voltage. The droop coefficient is... This is a non-linear exponent. Preferably, the value of n is 2 or 3.

[0091] When the distribution network voltage is normal, that is, when the voltage deviation is normal. When it is close to 0, The value is extremely small, and the local control unit keeps the charging and discharging plan unchanged.

[0092] When the voltage of the distribution network drops, for example, during peak evening electricity overload, the voltage deviation... Increase The increase is exponential, reducing the actual charging power. This method can quickly curb voltage collapse without requiring control commands from a local control unit or cloud-based optimization unit.

[0093] When communication is restored, in order to prevent multiple local control units in island defense mode from switching back to collective intelligence mode instantly, which would cause load changes and impact the distribution network, this application introduces Logistic chaotic mapping to generate chaotic sequence values ​​to delay the time for each local control unit to reconnect to the cloud optimization unit.

[0094] The value of the chaotic sequence is calculated as shown in formula (11): Formula (11); in, This represents the chaotic sequence value of the (k+1)th iteration of the current local control unit. Let be the chaotic sequence value of the k-th iteration. For the number of iterations, This is the growth rate parameter.

[0095] in, It is distributed between 0 and 1.

[0096] The reconnection waiting time is calculated as shown in formula (12): Formula (12); in, This is the reconnection waiting time, i.e., the waiting time for the current local control unit after communication is restored. For the current local control unit's number The chaotic sequence value of the next iteration. This is the preset maximum waiting time.

[0097] When communication recovers from a fault, to prevent multiple local control units from reconnecting simultaneously and causing secondary impacts on the distribution network, the local control units employ a soft-start mechanism based on chaotic sequences. This mechanism utilizes the inherent ergodicity and aperiodicity of chaotic sequences to generate unique and evenly distributed reconnection waiting times for each local control unit along the time axis, thereby dispersing concentrated reconnection loads in the time dimension and achieving smooth recovery of the distribution network load.

[0098] In response to communication disruptions or grid emergencies, such as when the edge controller detects a loss of heartbeat packets from the cloud or a severe voltage drop in the distribution network, it automatically blocks cluster scheduling prices from the cloud optimization unit and switches to islanding defense mode. In islanding defense mode, the controller executes nonlinear voltage droop control based on local voltage feedback, rapidly reducing charging power exponentially to prevent distribution transformer overload with millisecond-level response speed, ensuring the survivability of the power grid physical system.

[0099] The following describes a charging control method for electric vehicles in a power distribution network according to a specific embodiment. This specific embodiment is only for better illustrating the invention and does not constitute an undue limitation of the invention.

[0100] This embodiment is based on a simulation model of the IEEE 37-node distribution network system. The rated voltage of the distribution network system is set to 4.8 kV, and the connected loads include typical residential, commercial and industrial electricity consumption curves, which can realistically simulate the load distribution characteristics of the actual distribution network.

[0101] During the simulation, a simulated fleet of 1,000 electric vehicles was configured. This scale closely matches the actual scenario of large-scale electric vehicle access in a power distribution network, which is the standard simulation verification scale in this field. There is no need to build a physical experimental platform, and there are no high experimental costs.

[0102] These simulated vehicles are randomly connected to various nodes of the power distribution network, and the battery capacity of each electric vehicle is randomly distributed between 13.8 kWh and 24 kWh. In order to more realistically simulate the diversity of battery states in actual applications, the initial health state of all batteries is randomly set within a preset range to characterize different degrees of actual battery aging.

[0103] The simulation environment was set to a constant -10 degrees Celsius to simulate a typical winter cold scene in order to verify the applicability of the technical solution of this application in extreme environments.

[0104] To verify the effectiveness, superiority, and reliability of the technical solution of this application, four sets of comparative simulation schemes are set up in this embodiment, as shown in Table 1: Table 1 Comparison of Simulation Results Table 1 shows four sets of schemes. The first set is the benchmark scheme, which uses a disordered charging strategy. Simulated vehicles charge at maximum power immediately upon connecting to a charging station without any coordinated control measures. This is used to calibrate the distribution network operation without optimized control. The second set is the prior art comparison scheme, which uses a centralized optimization algorithm based on a static battery aging model for scheduling. This is used to compare the performance with the technical solution of this application, highlighting the innovative advantages of this application. The third set is the verification scheme, which applies the technical solution of this application, including constructing a battery aging cost model through online complex impedance measurement, collaboration between the cloud optimization unit and the local control unit, and an electrothermal coupling energy consumption model under low-temperature conditions, to verify the effectiveness of the core technology of this application. The fourth set is the fault scenario verification scheme, simulating a scenario of network-wide communication interruption during evening peak hours, to verify whether the islanding defense mode of this application can effectively maintain the stable operation of the distribution network and improve the reliability verification of the technical solution.

[0105] In Table 1, "N / A" indicates that it is not applicable. The second and third schemes have similar effects in reducing peak load on the distribution network, but their underlying mechanisms and long-term impacts differ fundamentally. The second scheme, using a static aging model, suffers from an inability of the scheduling algorithm to detect differences in the internal state of the batteries, resulting in approximately 15% of sub-healthy batteries with high internal resistance being incorrectly assigned high-frequency discharge tasks. While this scheduling optimizes the load curve in the short term, it accelerates the degradation of battery health and increases battery costs for users in the long term. In contrast, the third scheme can dynamically identify high-impedance batteries through real-time health factors and increase their virtual loss costs, allowing the optimization algorithm to automatically prioritize the allocation of adjustment tasks to healthy, low-impedance batteries. The results show that the scheme proposed in this application, while achieving peak shaving and valley filling, not only improves the utilization rate of healthy batteries but also provides protective rest scheduling for sub-healthy batteries, achieving Pareto optimal configuration of the entire network's battery asset lifespan.

[0106] Regarding communication requirements, the different solutions showed significant differences. The second solution required continuous uploading of raw data such as power, voltage, and state of charge for each vehicle, resulting in high daily communication traffic and placing a burden on communication networks in remote areas or with limited capacity. In contrast, the third solution only needed to transmit highly abstracted and aggregated schedulable feature vectors, reducing daily communication traffic by more than 98%. This optimization not only significantly reduced operating costs but also meant that the technology could be deployed on low-bandwidth networks such as narrowband IoT, greatly expanding its geographical applicability.

[0107] To verify survivability under extreme fault conditions, the fourth group simulated a scenario of a network-wide communication outage during the evening peak hours. Without a local defense mechanism, charging piles would either maintain the commands before the outage or resume disordered full charging, potentially causing transformer overload. The fourth group, employing the solution of this invention, triggered an islanding defense mode. When the communication outage caused a local voltage drop, the local control unit, based on a nonlinear droop control function, autonomously and rapidly reduced the charging power according to the real-time distribution network voltage. Simulation results show that the total system load was successfully limited within a safe threshold. Although slightly higher than the optimal scheduling value, it effectively maintained the safety baseline of the power distribution equipment capacity, demonstrating the protective role of the physical defense mechanism for the physical safety of the distribution network.

[0108] In simulations of extreme cold environments down to -10 degrees Celsius, the limitations of traditional solutions became apparent. Due to the limitations of the battery management system's low-temperature protection mechanism, the charging current in the first and second groups was restricted to a low level, resulting in slow charging and a poor user experience. The electrothermal coupling optimization strategy employed in the third group demonstrated significant advantages. By prioritizing power allocation for battery heating during the initial charging phase, the battery temperature quickly rose to its optimal operating range, allowing the system to safely apply a larger charging current and significantly shortening the total charging time. The results indicate that under extreme environmental conditions, active thermal management is a crucial element in overcoming charging bottlenecks and ensuring a superior user experience.

[0109] In summary, this invention systematically addresses the shortcomings of existing technologies in areas such as model distortion, communication dependence, and environmental adaptability through comprehensive innovation, encompassing precise battery status sensing, cloud-edge collaborative architecture, and safety defense under extreme operating conditions. Simulation data, from multiple dimensions including asset protection, communication efficiency, system robustness, and environmental adaptability, validates the beneficial effects of this invention and demonstrates its application potential in addressing the challenges of large-scale electric vehicle grid connection.

[0110] Figure 2 This application provides a schematic diagram of the structure of a charging control system for electric vehicles in a power distribution network, as shown in the embodiment of the present application. Figure 2 As shown, this application proposes a charging control system for electric vehicles in a power distribution network. The system 200 includes: a local control unit 210 and a cloud optimization unit 220. The local control unit is used to acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery status data, and battery complex impedance data; The local control unit is further configured to obtain the battery aging data of the current electric vehicle based on the battery complex impedance data; determine the real-time health factor of the current battery based on the battery aging data; calculate the battery loss cost based on the real-time health factor through a pre-built battery aging cost model; and obtain the schedulable feature vector of the current electric vehicle based on the battery loss cost, the user fuzzy charging demand data, and the current battery state data. The cloud optimization unit is used to determine the category cluster corresponding to the current electric vehicle based on the schedulable feature vector and through preset classification rules; based on the category cluster and the schedulable feature vector, with the optimization objective of minimizing the value of a pre-constructed target optimization function, solve the target optimization function to obtain the cluster scheduling price corresponding to the category cluster; The local control unit is used to receive the cluster scheduling price and, based on the cluster scheduling price and the schedulable feature vector, determine the charging and discharging plan of the current electric vehicle through a preset decision-making mechanism. The local control unit is used to execute the charging and discharging plan based on the current battery status data and the communication status between the local control unit and the cloud optimization unit.

[0111] It should be noted that the description of the above system embodiments is similar to the description of the above method embodiments, and has similar beneficial effects. For technical details not disclosed in the system embodiments of this application, please refer to the description of the method embodiments of this application for understanding.

[0112] It should be noted that, in the embodiments of this application, if the charging control method for electric vehicles in the power distribution network described above is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an electronic device to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0113] Correspondingly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements the steps in the charging control method for an electric vehicle in any of the above embodiments of the power distribution network. Correspondingly, embodiments of this application also provide a computer program product, which, when executed by a processor of an electronic device, is used to implement the steps in the charging control method for an electric vehicle in any of the above embodiments of the power distribution network.

[0114] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0115] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0116] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0117] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units; some or all of the units may be selected to achieve the purpose of the embodiments of this application according to actual needs. In addition, each functional unit in the embodiments of this application may be fully integrated into one processing unit, or each unit may be a separate unit, or two or more units may be integrated into one unit; the integrated unit may be implemented in hardware or in the form of hardware plus software functional units.

[0118] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause the device automatic test line to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0119] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined to obtain new method embodiments without conflict. The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined to obtain new method embodiments or device embodiments without conflict.

[0120] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A charging control method for electric vehicles in a power distribution network, characterized in that, A charging control system for electric vehicles applied in a power distribution network; the system includes: a local control unit and a cloud optimization unit; the method includes: The local control unit is used to acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery status data, and battery complex impedance data. Using the local control unit, based on the battery complex impedance data, the battery aging data of the current electric vehicle is obtained; based on the battery aging data, the real-time health factor of the current battery is determined; based on the real-time health factor, the battery loss cost is calculated through a pre-built battery aging cost model; based on the battery loss cost, the user fuzzy charging demand data, and the current battery state data, the schedulable feature vector of the current electric vehicle is obtained. Using the cloud optimization unit, based on the schedulable feature vector, the category cluster corresponding to the current electric vehicle is determined through preset classification rules; based on the category cluster and the schedulable feature vector, with the optimization objective being to minimize the value of a pre-constructed target optimization function, the target optimization function is solved to obtain the cluster scheduling price corresponding to the category cluster; Using the local control unit, the cluster scheduling price is received, and based on the cluster scheduling price and the schedulable feature vector, the charging and discharging plan of the current electric vehicle is determined through a preset decision-making mechanism; The charging and discharging plan is executed using the local control unit based on the current battery status data and the communication status between the local control unit and the cloud optimization unit.

2. The method according to claim 1, characterized in that, The step of determining the real-time health factor of the current battery based on the battery aging data includes: Using the local control unit, the standard value of the charge transfer impedance and the optimal operating temperature of the current battery are obtained; Using the local control unit, based on the battery complex impedance data, the battery aging data of the current battery is obtained by fitting a pre-built equivalent circuit model; the battery aging data includes the current charge transfer impedance and the current temperature. Using the local control unit, calculate the impedance aging term that shows the current charge transfer impedance deviating from the standard value of the charge transfer impedance; Using the local control unit, the temperature deviation term of the current temperature from the battery's optimal operating temperature is calculated; Using the local control unit, the impedance aging term and the temperature deviation term are fused and calculated through a nonlinear amplification mechanism to obtain the real-time health factor that generates an exponential penalty response when the impedance abnormally increases or the temperature deviates drastically.

3. The method according to claim 1, characterized in that, The method further includes: Using the local control unit, when the current battery establishes a charging connection with the local control unit, the DC-DC converter built into the local control unit is multiplexed as a signal generator. By performing micro-perturbation modulation on the pulse width modulation (PWM) signal of the power switching transistor in the DC-DC converter, a sinusoidal AC signal containing multiple different frequency components is superimposed on the DC charging current and applied to the current battery. Using the local control unit, the response data of the current battery to each frequency component is measured to obtain the corresponding voltage response signal and current response signal; Using the local control unit, the voltage response signal and the current response signal are denoised based on the discrete wavelet transform algorithm to obtain the denoised voltage response signal and the denoised current response signal. Using the local control unit, based on the denoised voltage response signal and the denoised current response signal, the complex impedance value corresponding to each frequency component is calculated to obtain the battery complex impedance data.

4. The method according to claim 1, characterized in that, The method further includes: Construct an initial battery aging model framework; Obtain historical battery aging data and historical health factors; The historical battery aging data and the historical health factors are input into the initial battery aging model framework to obtain battery aging parameters. Based on the battery aging parameters, the pre-built battery aging cost model is determined.

5. The method according to claim 1, characterized in that, The schedulable feature vector includes an urgency index, a flexibility index, a health confidence level, and a capacity contribution value; the schedulable feature vector of the current electric vehicle, obtained based on the battery loss cost, the user's fuzzy charging demand data, and the battery's current state data, includes: Obtain the rated power corresponding to the local control unit; Based on the user's fuzzy charging demand data and the battery's current state data, the urgency index and the flexibility index are obtained respectively; Based on the battery loss cost, determine the health confidence level that characterizes the battery's willingness to participate in discharge; Using the local control unit, a capacity contribution value is obtained based on the rated power and the current state data of the battery; Using the local control unit, based on the urgency index, the flexibility index, the health confidence level, and the capacity contribution value, the schedulable feature vector of the current electric vehicle is aggregated and generated.

6. The method according to claim 1, characterized in that, The process involves utilizing the cloud-based optimization unit to determine the category cluster corresponding to the current electric vehicle based on the schedulable feature vector and a preset classification rule; and then, based on the category cluster and the schedulable feature vector, solving the target optimization function with the objective of minimizing the value of a pre-constructed target optimization function to obtain the cluster scheduling price corresponding to the category cluster, including: Using the cloud optimization unit, the spatial distance between the schedulable feature vector and the center vector of each preset category cluster is calculated; Using the cloud optimization unit, the category cluster corresponding to the current electric vehicle is determined based on the spatial distance; Using the cloud optimization unit, the average urgency corresponding to the category cluster is determined based on the schedulable feature vector; The cloud-based optimization unit is used to obtain the real-time cost price of the power distribution network. Using the cloud optimization unit, the real-time cost electricity price and the average urgency are input into the pre-built target optimization function. Based on the horse herd optimization algorithm, the optimization objective is to minimize the value of the pre-built target optimization function. The target optimization function is then solved to obtain the cluster scheduling price corresponding to the category cluster and the power purchase capacity of the distribution network.

7. The method according to claim 1, characterized in that, The schedulable feature vector includes an urgency index; the charging / discharging plan includes a first charging / discharging plan and a second charging / discharging plan; the step of using the local control unit to receive the cluster scheduling price, and determining the current charging / discharging plan of the electric vehicle based on the cluster scheduling price and the schedulable feature vector through a preset decision-making mechanism includes: Using the local control unit, upon receiving the cluster scheduling price, the cluster scheduling price is compared with the urgency index and the battery wear cost, respectively; Using the local control unit, when the urgency index is greater than the cluster scheduling price, it is determined that rigid demand is dominant, and the first charge-discharge plan is determined to ignore economic costs and charge with the maximum allowable power. Using the local control unit, when the urgency index is less than or equal to the cluster scheduling price, it is determined that economy is the primary factor, and a second charge-discharge plan is determined to reduce the charging power in response to the adjustment signal; the charging priority of the second charge-discharge plan is higher than that of the first charge-discharge plan.

8. The method according to claim 1, characterized in that, The current battery status data includes the battery temperature; the step of executing the charge / discharge plan using the local control unit, based on the current battery status data and the communication status between the local control unit and the cloud optimization unit, includes: Using the local control unit, when the battery temperature is lower than a preset temperature threshold, the heating power and charging power of the current battery are allocated based on a preset battery thermal model to preheat the current battery. After the current battery has been preheated, the local control unit executes the charge / discharge plan based on the communication status between the local control unit and the cloud optimization unit.

9. The method according to claim 8, characterized in that, The step of executing the charging and discharging plan based on the communication status between the local control unit and the cloud optimization unit includes: When communication is normal, the charging and discharging plan is executed using the local control unit; When communication is interrupted, the local control unit is used to determine an updated charging and discharging plan based on the power distribution network data; the updated charging and discharging plan includes reducing the charging power corresponding to the charging and discharging plan and keeping the charging power corresponding to the charging and discharging plan unchanged. When communication is restored, the local control unit connects to the cloud optimization unit after the preset communication quality conditions are met.

10. A charging control system for electric vehicles in a power distribution network, characterized in that, The system includes: a local control unit and a cloud optimization unit; The local control unit is used to acquire power distribution network data, current user fuzzy charging demand data for electric vehicles, current battery status data, and battery complex impedance data; The local control unit is further configured to obtain the battery aging data of the current electric vehicle based on the battery complex impedance data; determine the real-time health factor of the current battery based on the battery aging data; calculate the battery loss cost based on the real-time health factor through a pre-built battery aging cost model; and obtain the schedulable feature vector of the current electric vehicle based on the battery loss cost, the user fuzzy charging demand data, and the current battery state data. The cloud optimization unit is used to determine the category cluster corresponding to the current electric vehicle based on the schedulable feature vector and through preset classification rules; based on the category cluster and the schedulable feature vector, with the optimization objective of minimizing the value of a pre-constructed target optimization function, solve the target optimization function to obtain the cluster scheduling price corresponding to the category cluster; The local control unit is used to receive the cluster scheduling price and, based on the cluster scheduling price and the schedulable feature vector, determine the charging and discharging plan of the current electric vehicle through a preset decision-making mechanism. The local control unit is used to execute the charging and discharging plan based on the current battery status data and the communication status between the local control unit and the cloud optimization unit.

Images