Intelligent charging priority distribution system and method based on multi-device identification

Abstract

The invention discloses an intelligent charging priority distribution system and method based on multi-device identification, and particularly relates to the technical field of charging priority distribution. Equipment identity and charging demand information are obtained through an intelligent identification technology, a priority score is calculated in combination with a preset weight parameter, the real-time performance of priority distribution is ensured by adopting a dynamic threshold adjustment strategy, power demand fluctuation in a short time in the future is predicted based on a load prediction algorithm of machine learning, power distribution is optimized in advance, and the power distribution efficiency is improved. The method comprises the steps of reducing instantaneous load impact, monitoring equipment power in real time in the charging process, comparing the equipment power with the maximum distributable power, and dynamically adjusting the charging power or switching the charging sequence through a self-adaptive power adjustment mechanism if abnormality is detected, so as to guarantee the stability of a power grid and the safety of the equipment, and the method can effectively improve the utilization efficiency of charging resources and reduce the energy consumption. Overload of a power grid is prevented, multi-device charging scheduling is optimized, and safety, reliability and intelligence of the charging process are ensured.

Description

Technical Field

[0001] This invention relates to the field of charging priority allocation technology, and more specifically to an intelligent charging priority allocation system and method based on multi-device identification. Background Technology

[0002] Intelligent charging priority allocation based on multi-device identification refers to the process in systems supporting simultaneous charging of multiple devices. This involves using intelligent identification technologies (such as RFID, NFC, Bluetooth, Wi-Fi, and AI algorithms) to acquire device identity, battery status, and charging needs. Combined with Dynamic Load Management (DLM) algorithms, charging resources are automatically allocated based on factors such as battery level, device type, and user permissions. This method optimizes charging efficiency, improves energy utilization, and ensures that critical or urgently needed devices receive priority charging, thereby enhancing system intelligence and user experience.

[0003] In existing technologies, static or simple dynamic load management strategies allocate charging resources according to fixed rules or real-time power demand when multiple devices are charging. However, when multiple high-power devices (such as electric buses and industrial equipment) are simultaneously connected to the charging system within a short period of time and are incorrectly assigned high priority by the system, the DLM (Dynamic Load Management System) may fail to adjust power allocation in time, causing the total load to momentarily exceed the grid's capacity. Furthermore, if the charging power demand forecast is incorrect, it may lead to delayed power allocation, damaging charging equipment and even posing a fire risk. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent charging priority allocation system and method based on multi-device identification, so as to solve the shortcomings of the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent charging priority allocation method based on multi-device identification, comprising: The system uses intelligent identification technology to obtain the unique identification information of multiple devices connected to the charging system, collects the current power level, battery health status, maximum supported charging power and historical charging behavior data of each device, and generates device charging demand information. Based on the charging demand information of the devices and combined with the preset weight parameters, the priority score of each device is calculated, and a dynamic threshold adjustment strategy is adopted to dynamically update the priority calculation model when the number of charging devices and the total power change. A machine learning-based load forecasting algorithm is used to predict power demand fluctuations in the near future by combining equipment power demand and grid load conditions. Based on the forecast results, the maximum allocable power of each device is adjusted in advance. During the charging process, the actual charging power of the device is monitored in real time and compared with the maximum allocable power. If an abnormal power allocation is detected, the charging power of different devices is dynamically adjusted or the charging order is dynamically switched based on priority through an adaptive power adjustment mechanism.

[0006] Preferably, the intelligent identification technology includes wired identification and wireless identification to ensure that each device has unique identification information.

[0007] Preferably, the preset weight parameters are used to calculate the device priority score, including but not limited to current battery level, battery health status, maximum supported charging power, user permissions, urgency level, and historical charging behavior. Devices with battery levels below 10% have their priority automatically increased, while devices with battery levels above 80% have their priority decreased.

[0008] Preferably, the dynamic threshold adjustment strategy is optimized based on the number of charging devices and the total power load. If a new device is added, the priority score of all devices is recalculated. If a device is disconnected, the power resources occupied by that device are released and redistributed to the remaining devices.

[0009] Preferably, a power demand anomaly index is generated after analyzing the equipment power demand. The method for obtaining the power demand anomaly index is as follows: First, power demand data over a period of time is collected from the charging management system. From the historical power demand data, sorted in ascending order, the first quartile (Q1) is calculated: the value at the 25th percentile of the data; the third quartile (Q3) is the value at the 75th percentile. IQR, or interquartile range, represents the range between Q1 and Q3. Anomaly thresholds are set. For detected abnormal power demand, a power demand anomaly index is calculated to measure its degree of abnormality: if the power demand falls within the normal range, the anomaly index is set to 0; if the power demand exceeds the upper or lower limit, the anomaly index is calculated based on the degree of deviation. The expression is: In the formula, This is the power demand anomaly index. This represents the actual power requirements of the current equipment. These are boundary values ​​within the normal range.

[0010] Preferably, the total power consumption of all current devices is monitored, and a power grid load fluctuation index is generated by analyzing the power grid load change trend in conjunction with historical data. The method for obtaining the power grid load fluctuation index is as follows: First, collect the total power consumption data of the power grid to form a time series. Then, use Fourier transform to convert the time series data of the power grid load to the frequency domain to obtain the frequency distribution of the load power. Calculate the load's spectral information to obtain the amplitude and energy corresponding to each frequency component. Distinguish between low-frequency and high-frequency components, calculate the proportion of high-frequency power fluctuations in the total load, i.e., calculate the energy of each frequency component, and obtain the load fluctuation amplitude in different frequency ranges. Based on the proportion of high-frequency fluctuation energy, define the power grid load fluctuation index, expressed as: ; This refers to the energy of high-frequency components, specifically the total energy of frequency components in the power grid load signal that are above a threshold. The total energy refers to the total power spectral energy of the entire load signal. This is the power grid load fluctuation index.

[0011] Preferably, the power demand anomaly index and the grid load fluctuation index are converted into a comprehensive feature vector. The comprehensive feature vector is used as the input to the machine learning model. The machine learning model uses the prediction of the power demand fluctuation value label in the short term for each set of comprehensive feature vectors as the prediction objective and minimizes the sum of prediction errors for all power demand fluctuation value labels in the short term as the training objective. The machine learning model is trained until the sum of prediction errors converges and the model training stops. The power demand fluctuation value in the short term is determined based on the model output. The machine learning model is a multinomial regression model.

[0012] Preferably, the obtained power demand fluctuation value in the near future is compared with a preset threshold. If the power demand fluctuation value in the near future is greater than or equal to the preset threshold, it indicates that the power demand fluctuation in the near future is large, and load management measures need to be taken to reduce the instantaneous power fluctuation of the power grid. If the power demand fluctuation value in the near future is less than the preset threshold, it indicates that the power demand fluctuation in the near future is small, and charging efficiency can be improved and power allocation optimized.

[0013] Preferably, the adaptive power adjustment mechanism includes calculating the power allocation deviation: ; Indicates the actual power of the device Power allocation with the system The deviation between them The actual charging power, i.e., the charging power currently measured by the device. The maximum allocatable power is the maximum charging power that the system sets for this device. Set a deviation threshold ϵ, if If the power is normal, no adjustment is needed; if If the power exceeds the allocated value, the charging power needs to be reduced; if If the power is lower than the allocated value, the charging power needs to be increased.

[0014] The present invention also provides an intelligent charging priority allocation system based on multi-device identification, including a data acquisition module, a priority score calculation module, a power prediction module and a priority adjustment module; Data acquisition module: It uses intelligent identification technology to obtain the unique identification information of multiple devices connected to the charging system, collects the current power, battery health status, maximum supported charging power and historical charging behavior data of each device, and generates device charging demand information; Priority score calculation module: Based on the charging demand information of the devices and combined with preset weight parameters, calculate the priority score of each device, and adopt a dynamic threshold adjustment strategy to dynamically update the priority calculation model when the number of charging devices and the total power change. Power prediction module: It adopts a load prediction algorithm based on machine learning, combined with the power demand of the equipment and the power grid load, to predict the fluctuation of power demand in the near future, and adjust the maximum allocable power of each device in advance based on the prediction results. Priority adjustment module: During the charging process, the actual charging power of the device is monitored in real time and compared with the maximum allocable power. If an abnormal power allocation is detected, the charging power of different devices is dynamically adjusted or the charging order is dynamically switched based on priority through an adaptive power adjustment mechanism.

[0015] The technical effects and advantages provided by the present invention in the above technical solution are as follows: 1. This invention achieves efficient management of the charging system through technologies such as intelligent identification, dynamic priority calculation, machine learning prediction, and adaptive power adjustment. Compared to traditional static or simple dynamic load management, this invention can accurately identify each charging device, collect its power level, battery health status, maximum supported charging power, and historical charging behavior, and calculate priorities based on preset weight parameters to ensure that emergency devices (such as police cars, ambulances, and UPS) are charged first. Furthermore, a dynamic threshold adjustment strategy is adopted to automatically update the priority calculation model when devices are connected or disconnected, optimizing power allocation, improving resource utilization, and preventing low-priority devices from being unable to receive charging opportunities for extended periods.

[0016] 2. This invention further introduces a machine learning-based load prediction algorithm, combining a power demand anomaly index and a grid load fluctuation index to accurately predict power demand fluctuations in the near future, thereby optimizing charging power allocation in advance, reducing grid load impacts, and improving system stability. During charging, the deviation between the actual charging power and the maximum allocable power of the equipment is monitored in real time, and dynamic adjustments are made through an adaptive power adjustment mechanism to ensure reasonable power allocation and prevent equipment overload or power waste. This method not only improves charging efficiency but also effectively reduces the risk of equipment damage or even fire caused by power allocation lag. Simultaneously, it supports smart grid optimization scheduling, reducing the impact of peak loads on the grid, making the charging system safer, more efficient, and more intelligent. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0018] Figure 1 This is a flowchart of the method of the present invention.

[0019] Figure 2 This is a system module diagram of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1, please refer to Figure 1 As shown in this embodiment, the intelligent charging priority allocation method based on multi-device identification includes: The system uses intelligent identification technology to obtain the unique identification information of multiple devices connected to the charging system, collects the current power level, battery health status, maximum supported charging power and historical charging behavior data of each device, and generates device charging demand information. Based on the charging demand information of the devices and combined with the preset weight parameters, the priority score of each device is calculated, and a dynamic threshold adjustment strategy is adopted to dynamically update the priority calculation model when the number of charging devices and the total power change. A machine learning-based load forecasting algorithm is used to predict power demand fluctuations in the near future by combining equipment power demand and grid load conditions. Based on the forecast results, the maximum allocable power of each device is adjusted in advance. During the charging process, the actual charging power of the device is monitored in real time and compared with the maximum allocable power. If an abnormal power allocation is detected, the charging power of different devices is dynamically adjusted or the charging order is dynamically switched based on priority through an adaptive power adjustment mechanism.

[0022] To accurately manage multiple connected charging devices, the system first needs to obtain unique identification information to ensure accurate differentiation of different devices and prevent duplicate identification or identity forgery. The main identification technologies and data acquisition methods are as follows: Wired-connection-based device identification: suitable for USB charging devices (such as smartphones, tablets, and laptops), as well as DC fast charging stations for electric vehicles.

[0023] USB port identification: The device's VID (Manufacturer ID), PID (Product ID), and SN (Serial Number) are read via the USB Power Delivery (USB PD) protocol as unique identifiers. Electric vehicle charging gun identification (CCS, CHAdeMO, GB / T standards): The EVCC ID (Electric Vehicle Controller ID) transmitted from the vehicle's BMS (Battery Management System) is read to ensure each vehicle is uniquely identifiable. This is further verified by combining the license plate number / charging card information.

[0024] Device identification based on wireless communication: Suitable for wireless charging, smart home devices, and shared charging facilities, using short-range or long-range communication for identification; RFID / NFC (Near Field Communication): Suitable for wireless charging devices (such as smartphones, smartwatches, and wireless earphones). It reads the device's UID (Unique Identifier) ​​via NFC tags or RFID chips and matches it with a database to obtain the device's identity. Bluetooth (Bluetooth Low Energy, BLE): Suitable for Bluetooth charging devices (such as electric scooters and smart wearable devices). It reads the device's MAC address and signal strength (RSSI), determines whether the device is within charging range based on RSSI changes, and matches historical device data. Wi-Fi / Cloud identification: Suitable for smart charging piles and IoT devices, it identifies devices through Wi-Fi authentication (such as MAC address + DHCP-assigned IP) and uploads device information to a cloud management system.

[0025] To prevent device forgery or identity spoofing, multi-factor authentication (MFA) is employed, such as: device ID + user account (APP login, IC card, fingerprint, etc.). Device ID + charging behavior history comparison (abnormal behavior detection, such as a device suddenly consuming abnormally high power).

[0026] After identifying the device, the system needs to obtain the device's charging requirements, including current battery level, battery health status, maximum supported charging power, and historical charging behavior, in order to optimize intelligent charging strategies.

[0027] Current battery information collection: Used to determine the device's remaining battery power and decide charging priority: Smartphones / Laptops: Obtain battery power (unit: %) via USB PD communication protocol. Remaining battery power is read via Qi wireless charging protocol; some devices support feedback on current power demand. Electric Vehicles / High-Power Devices: Obtain SoC (State of Charge) data (unit: %) via BMS (Battery Management System) + CAN bus communication. Combined with GPS, predict the vehicle's remaining driving range (e.g., prioritize charging if less than 50 km).

[0028] To assess whether a battery requires a special charging mode (such as low-temperature protection or slow charging) to avoid battery damage: Smart devices (phones, laptops): Read the Battery Health Score via the Android / iOS Battery API. Electric vehicles: Read the State of Health (SOH) via the OBD-II interface or CAN bus. If the SOH < 80%, the system may prioritize a low-speed charging mode to extend battery life.

[0029] Different devices support different maximum power outputs, and the system needs to obtain this data to optimize charging allocation: Smartphones / Wireless Charging: Determine the maximum power via USB PD (maximum power 5W-100W) or Qi wireless charging protocol (5W, 10W, 15W). Electric Vehicles / High-Power Devices: Read the maximum output power allowed by the charging station's EVSE (Electric Vehicle Supply Equipment) (e.g., 22kW, 50kW, 350kW). Read BMS communication to determine the vehicle's maximum acceptable power and prevent overload.

[0030] Historical charging behavior data is used to optimize charging strategies and improve intelligent allocation efficiency: Charging time statistics (user's daily charging time). Charging frequency (whether the user is a high-frequency user, such as a taxi driver). Peak power demand (whether fast charging is frequently used). Whether there is abnormal charging behavior (such as frequent disconnections and reconnections, which may be malicious priority boosting).

[0031] Based on the above data, the system generates device charging demand information, including: device unique identifier (ID), current battery level (SoC %), battery health status (SOH %), maximum supported charging power (W), historical charging behavior (time, frequency, fast charging habits), user category (regular user / VIP user / emergency vehicle), and environmental factors (such as adjusting charging strategy when the temperature is too high). This data will be used for subsequent charging priority calculation and dynamic power allocation.

[0032] The system collects the following key data from the connected charging devices: Current battery level (SoC, State of Charge): Used to determine the device's remaining battery level.

[0033] Battery State of Health (SoH): Ensures the battery is in good condition and determines whether to enable protective charging strategies.

[0034] Maximum supported charging power: Used to allocate appropriate charging power to prevent overload or inefficient charging.

[0035] Device Type: Devices such as electric vehicles, smartphones, and servers may have different priority weights. User Permissions: The system determines whether the user owns the device is a VIP, a regular user, or a public service device, adjusting priority accordingly. Historical Charging Behavior: Records the device's charging frequency, charging time, and whether it has been chronically low on battery, optimizing charging strategies. Urgency Level: Emergency vehicles (ambulances, police cars), critical equipment (UPS power servers), etc., require automatic priority increases. The system sets preset weights based on the importance of different parameters for subsequent priority score calculations.

[0036] The system combines the collected data and the set weight parameters to calculate an initial priority score for each device, specifically including the following logic: Low-battery devices are given priority: if a device's battery level is less than 10%, it is given extremely high weight; if a device's battery level is greater than 80%, its priority is reduced to minimize unnecessary power allocation.

[0037] Equipment type affects priority: Medical equipment, emergency service vehicles, and data center UPS equipment are given higher priority; ordinary consumer devices (such as mobile phones and laptops) have relatively lower priority.

[0038] Historical charging behavior optimization: If a device is in a low battery state for a long time, it may indicate that the user has a high charging demand, and the system will appropriately increase the priority; if the device charges frequently but does not actually consume electricity, the priority may be appropriately reduced to prevent resource waste.

[0039] Charging demand and power matching: If the current power grid allows the device to provide its maximum supported charging power, the priority can be appropriately increased; if the power grid load is high, it may be necessary to reduce the charging power of some devices to avoid overload.

[0040] After the calculation is completed, the system generates an initial priority score for each device, which is then used for the next step of dynamic adjustment.

[0041] Since the number of charging devices and the total power of the grid change over time, the system needs to dynamically update the priority calculation model to ensure reasonable allocation of charging resources.

[0042] If a new device is connected to the charger: calculate the priority score of the new device and determine whether it is necessary to reallocate the charging power of existing devices; if the current number of devices is close to the maximum power supply capacity, reduce the charging power of some low-priority devices to avoid grid overload.

[0043] If a device disconnects from charging: release the charging resources occupied by that device and re-evaluate the priority scores of the remaining devices; if the released power is large, the charging power of other devices can be appropriately increased to improve charging efficiency.

[0044] If the total power load is too high: reduce the power allocation of low-priority devices to ensure that high-priority devices receive sufficient charging resources; or suspend the charging of some low-priority devices or switch to a low-power charging mode (such as trickle charging).

[0045] If the total power load is low: allow low-priority devices to accelerate charging to improve charging efficiency; allocate more power to fast charging devices to improve the overall charging completion rate.

[0046] The priority calculation formula is optimized based on short-term charging data: If a device consistently maintains a high priority across multiple charging attempts, the system will perform behavioral analysis to determine whether it truly needs high-priority charging, preventing resource abuse. If a device remains at a low priority for an extended period without receiving charging, the system will appropriately increase its priority to avoid "charging starvation."

[0047] AI is used to predict future load conditions: if AI predicts that a large number of devices will be connected in the next few minutes, the charging power allocation will be adjusted in advance to avoid sudden grid overload; if AI predicts that some devices are about to finish charging, power will be reserved for other devices in advance to improve resource utilization.

[0048] Update charging priorities and execute power allocation: The system rearranges the charging order of devices based on the latest calculated priority score; adjusts the charging power allocation strategy: high-priority devices receive higher charging power (such as fast charging mode). Low-priority devices may be allocated lower power or even enter a queuing state. Continuously monitor the charging process; if an anomaly is detected (such as power overload, abnormal device disconnection, etc.), the system recalculates priorities and dynamically adjusts the charging power.

[0049] Device power requirements are influenced by various factors, including current battery level, battery health, historical charging behavior, and device type. When predicting future power requirements, the system first calculates the theoretical demand based on the device's current system charge (SoC) and the battery's maximum charging power; for example, a low-battery device may require higher charging power. Secondly, it identifies the device's charging habits by considering historical charging behavior; for instance, some users prefer fast charging, while some devices may frequently use low-power slow charging. Furthermore, the system also considers the battery's state of health (SoH), as aging batteries may require slower charging modes to extend their lifespan. Machine learning models analyze this data to predict the device's charging trends in the near future and adjust charging power in advance to avoid power imbalances.

[0050] After analyzing the power demand of the equipment, a power demand anomaly index is generated. The method for obtaining the power demand anomaly index is as follows: First, power demand data over a period of time is collected from the charging management system, such as the actual charging power of each device within the last 7 days, 30 days, or a specific time window (e.g., the past hour). The data may include: device ID (used to uniquely identify the charging device), charging time (timestamp, used to analyze time trends), actual charging power (unit: W or kW), charging mode (fast charging / slow charging / trickle charging), and device category (smartphone, laptop, electric vehicle, etc.). From historical power demand data, arranged in ascending order, calculate: the first quartile Q1: the value at the 25th percentile of the data, representing a typical value for lower power demand; the third quartile Q3: the value at the 75th percentile of the data, representing a typical value for higher power demand. Q1 and Q3 can be used to measure the distribution of data, helping to detect abnormal fluctuations in power demand. IQR, or interquartile range, represents the numerical range between Q1 and Q3. IQR represents the primary range of data distribution; data should typically be concentrated within this range. If a device's charging power exceeds the normal IQR range, it may indicate an anomaly.

[0051] Use IQR to calculate the normal range of power demand and set the judgment boundaries for outliers: Lower Bound = Q1 - 1.5 × IQR; Upper Bound = Q3 + 1.5 × IQR. If the charging power of a device is lower than the lower bound, it indicates that the device's power demand is abnormally low (possibly due to device malfunction, disconnection from charging, etc.). If it is higher than the upper bound, it indicates that the device's power demand is abnormally high (possibly due to abnormal device discharge, abnormal fast charging, etc.).

[0052] For detected abnormal power demand, a power demand anomaly index is calculated to measure the degree of abnormality: if the power demand falls within the normal range (between Q1 and Q3), the anomaly index is set to 0 (indicating normal). If the power demand exceeds the upper or lower limit, the anomaly index is calculated according to the degree of exceeding the range, expressed as: In the formula, This is the power demand anomaly index. This represents the actual power requirements of the current equipment. These are the boundary values ​​within the normal range; that is, when the power is too high, the upper limit is used, and when the power is too low, the lower limit is used. The larger the portion exceeding the upper limit, the higher the anomaly index. The larger the portion below the lower limit, the higher the anomaly index.

[0053] The anomaly index can be used to prioritize abnormal situations. For example, if device A has a power requirement of 40.0kW, while the highest normal range is 25.0kW, then its anomaly index is relatively high. If device C has a power requirement of 90.0kW, while the highest normal range is 55.0kW, then its anomaly index is even higher, and its priority is higher.

[0054] When the system detects devices with high anomaly indices, the following dynamic adjustment measures can be taken: reduce the charging power of abnormally high-power devices to avoid overloading the power grid. Increase the charging power of abnormally low-power devices to check if the failure to charge is due to a fault. Optimize threshold settings based on historical data to ensure that the IQR detection method can adapt to the power requirements of different device types.

[0055] The grid load is affected by factors such as the total power demand of current charging equipment, power supply status, peak-valley pricing strategies, and weather conditions. When predicting power fluctuations in the near future, the system needs to monitor the total power consumption of all current equipment in real time and analyze grid load trends using historical data. For example, during peak electricity consumption periods (such as the evening peak from 6:00 PM to 10:00 PM), the grid load is high, and the system may need to reduce the charging power of some non-emergency devices to prevent overload. Simultaneously, weather factors (such as increased air conditioning load due to high temperatures) or unforeseen events (such as large-scale equipment access) can also affect the grid's power supply capacity. Machine learning models analyze these factors to predict future grid load levels and optimize power allocation strategies in advance, ensuring that charging demand is met without impacting grid stability.

[0056] The total power consumption of all current devices is monitored, and a power grid load fluctuation index is generated by analyzing the power grid load change trend in conjunction with historical data. The method for obtaining the power grid load fluctuation index is as follows: First, collect total power consumption data from the power grid to form a time series. Data sources include: real-time power monitoring systems (such as smart meters, smart charging piles, and power grid SCADA systems); historical power data from charging equipment (such as electric vehicles, data centers, and industrial loads); and power grid frequency data (such as 50Hz / 60Hz power grid fluctuations). Use Fourier transform to convert the time series data of the power grid load to the frequency domain to obtain the frequency distribution of the load power. Calculate the load's spectral information to obtain the amplitude and energy corresponding to each frequency component. Distinguish between low-frequency and high-frequency components: Low frequency (0-0.01Hz): represents long-cycle load changes, such as daytime / nighttime load fluctuations. Mid-frequency (0.01-0.1Hz): represents regular load fluctuations, such as office hours vs. rest hours. High frequency (>0.1Hz): represents short-term power fluctuations, such as frequent device connection / disconnection. To measure the degree of power grid load fluctuation, it is necessary to calculate the proportion of high-frequency power fluctuations in the total load, i.e., calculate the energy (power spectral density) of each frequency component to obtain the load fluctuation amplitude in different frequency ranges. Set thresholds to distinguish between normal and abnormal load fluctuations: if the proportion of high-frequency fluctuation energy is below a certain threshold (e.g., 5%), it indicates that the load is relatively stable. If the proportion of high-frequency fluctuation energy exceeds the threshold (e.g., 20%), it indicates that the grid load fluctuates significantly and there may be a problem.

[0057] Based on the proportion of high-frequency fluctuation energy, the power grid load fluctuation index PGLFI is defined as follows: ; This refers to the energy of high-frequency components, specifically the total energy of frequency components in the power grid load signal that are above a threshold. The total energy refers to the total power spectral energy of the entire load signal. The PGLFI (Power Grid Load Fluctuation Index) is used to quantify the stability of the power grid load: a lower PGLFI (e.g., close to 0) indicates a stable power grid load with minimal fluctuations. A moderate PGLFI (e.g., 0.2-0.5) indicates some load fluctuations, but these are within the normal range. An excessively high PGLFI (e.g., >0.7) indicates severe power grid load fluctuations, which may affect system stability.

[0058] After calculating the PGLFI, the system can intelligently adjust and optimize the charging management strategy: Low PGLFI (stable load): Allows all devices to charge normally without load adjustment. The charging power of fast-charging devices can be appropriately increased to improve charging efficiency.

[0059] PGLFI Moderate (Normal Fluctuations): Smoothly distributes power to some devices to reduce peak loads. It combines machine learning to predict short-term power demand and optimize the allocation strategy.

[0060] PGLFI too high (severe load fluctuations): Reduce the charging power of some non-urgent devices to reduce load fluctuations. In extreme cases, dynamic load shedding can be used to limit the current or suspend charging of low-priority devices.

[0061] The power demand anomaly index and grid load fluctuation index are converted into a comprehensive feature vector. This comprehensive feature vector is then used as input to a machine learning model. The machine learning model uses the prediction of the power demand fluctuation value label in the next short period of time for each set of comprehensive feature vectors as its prediction objective. The training objective is to minimize the sum of prediction errors for all power demand fluctuation value labels in the next short period of time. The machine learning model is trained until the sum of prediction errors converges, at which point the model training stops. The power demand fluctuation value in the next short period of time is determined based on the model output. The machine learning model is a multinomial regression model.

[0062] The method for obtaining the power demand fluctuation value in the near future is as follows: Obtain the corresponding function expression from the comprehensive feature vector training data of the trained machine learning model. In the formula, It is the output function of the model. This is the power demand anomaly index. This is the power grid load fluctuation index. This represents the expected fluctuation in power demand in the near future.

[0063] The obtained power demand fluctuation value in the near future is compared with a preset threshold. If the power demand fluctuation value in the near future is greater than or equal to the preset threshold, it means that the power demand fluctuation in the near future (e.g., 5-10 minutes) is too large, which may lead to grid load instability or even overload risk. Load management measures need to be taken to reduce the instantaneous power fluctuation of the grid.

[0064] Adjustment strategies include: reducing charging power for low-priority devices: for example, current limiting for non-urgent devices (such as ordinary electric vehicles and office equipment) to reduce instantaneous power demand. Proportional adjustment strategies can be adopted, such as reducing power demand by 10%-30% based on the current proportion. Delaying charging time for some devices: postponing charging times for low-priority or non-immediately-needed devices (such as vehicles charging at night) to alleviate peak load. Smart off-grid charging: if there are multiple charging stations or power supply points, dynamically adjusting the charging load of each station to balance the power distribution of the entire grid. Activating energy storage systems (such as batteries or supercapacitors) for auxiliary power supply: if the smart grid is equipped with energy storage devices, the energy storage system can be activated during periods of high power demand to smooth power fluctuations.

[0065] If the power demand fluctuation in the near future is less than a pre-set threshold, it indicates that the power demand fluctuation is relatively small and the power grid is operating relatively smoothly. Without affecting grid stability, charging efficiency can be improved and power distribution optimized.

[0066] Adjustment strategies include: increasing charging power for high-priority devices: for example, appropriately increasing charging power for emergency vehicles (ambulances, police cars) and data center UPS to speed up the charging process. Allowing more devices to charge simultaneously: increasing the number of parallel charging devices when power load is low to improve overall charging utilization. Utilizing low-electricity-price periods for intelligent supplementary charging: if the system is connected to a smart grid (such as peak-valley pricing strategies), charging can be done in advance during low-load periods to reduce load pressure during future peak periods.

[0067] During charging, the actual charging power of the equipment needs to be monitored in real time and compared with the maximum allocable power. When the allocated power is abnormal (such as too high, too low, or fluctuating drastically), an adaptive power adjustment mechanism is used to dynamically adjust the power allocation to optimize power distribution and ensure grid stability.

[0068] Calculate power distribution deviation: ; Indicates the actual power of the device Power allocation with the system If the deviation is large, it indicates an abnormality in the charging power distribution. The actual charging power is the charging power currently measured by the device (unit: kW). Maximum allocable power, which is the maximum charging power (unit: kW) set by the system for this device.

[0069] Set a deviation threshold ϵ, if If the power is normal, no adjustment is needed; if If the power exceeds the allocated value, the charging power needs to be reduced; if If the power is lower than the allocated value, the charging power needs to be increased.

[0070] The new power allocation value (dynamically adjusted) is calculated using the following expression: ; The adjusted charging power is the newly calculated charging power of the device, used to correct power anomalies. α is an adjustment coefficient used to control the power adjustment rate, with a value range of 0 < α ≤ 1, used to prevent excessively rapid adjustments that could cause system oscillations. When , The impact on the power grid is reduced by decreasing the impact of the power grid; when , Increase and improve charging efficiency.

[0071] The expression for calculating device priority changes and adjusting charging order is: In the formula, Device priority score: used to dynamically adjust the charging order; the higher the score, the higher the priority for charging. U represents user priority: such as VIP users, key grid equipment, and electric buses have higher priority. SoC represents current battery status: the lower the battery (smaller SoC), the higher the priority. Remaining charging time: Calculates the estimated charging time for each device, prioritizing devices with shorter charging times. For power allocation error: If the charging power of the device fluctuates greatly, its priority can be appropriately reduced to avoid system instability. Weighting parameters: used to adjust the degree of influence of different factors, and can be optimized according to system requirements.

[0072] If the newly calculated device priority score If the device priority score is greater than or equal to the initial device priority score, the device charging order is increased; if the newly calculated device priority score is... If a device's priority score is lower than the initial priority score, its charging order will be reduced. High-priority devices will be charged first (e.g., devices with low battery or those in urgent situations). Low-priority devices may be put into a queue or charged at reduced power to ensure system stability.

[0073] Example 2, please refer to Figure 2As shown, the intelligent charging priority allocation system based on multi-device identification described in this embodiment includes a data acquisition module, a priority score calculation module, a power prediction module, and a priority adjustment module. Data acquisition module: It uses intelligent identification technology to obtain the unique identification information of multiple devices connected to the charging system, collects the current power, battery health status, maximum supported charging power and historical charging behavior data of each device, and generates device charging demand information; Priority score calculation module: Based on the charging demand information of the devices and combined with preset weight parameters, calculate the priority score of each device, and adopt a dynamic threshold adjustment strategy to dynamically update the priority calculation model when the number of charging devices and the total power change. Power prediction module: It adopts a load prediction algorithm based on machine learning, combined with the power demand of the equipment and the power grid load, to predict the fluctuation of power demand in the near future, and adjust the maximum allocable power of each device in advance based on the prediction results. Priority adjustment module: During the charging process, the actual charging power of the device is monitored in real time and compared with the maximum allocable power. If an abnormal power allocation is detected, the charging power of different devices is dynamically adjusted or the charging order is dynamically switched based on priority through an adaptive power adjustment mechanism.

[0074] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0075] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0076] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0077] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes 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.

Claims

1. A method for intelligent charging priority allocation based on multi-device identification, characterized in that: include: The system uses intelligent identification technology to obtain the unique identification information of multiple devices connected to the charging system, collects the current power level, battery health status, maximum supported charging power and historical charging behavior data of each device, and generates device charging demand information. Based on the charging demand information of the devices and combined with the preset weight parameters, the priority score of each device is calculated, and a dynamic threshold adjustment strategy is adopted to dynamically update the priority calculation model when the number of charging devices and the total power change. A machine learning-based load forecasting algorithm is used to predict power demand fluctuations in the near future by combining equipment power demand and grid load conditions. Based on the forecast results, the maximum allocable power of each device is adjusted in advance. During the charging process, the actual charging power of the device is monitored in real time and compared with the maximum allocable power. If an abnormal power allocation is detected, the charging power of different devices is dynamically adjusted or the charging order is dynamically switched based on priority through an adaptive power adjustment mechanism. 2.The multi-device identification based intelligent charging priority allocation method according to claim 1, characterized in that: The intelligent identification technology includes wired and wireless identification, which is used to ensure that each device has unique identification information. 3.The multi-device identification based intelligent charging priority allocation method of claim 1, wherein: The preset weight parameters are used to calculate the device priority score, including but not limited to current battery level, battery health status, maximum supported charging power, user permissions, urgency level, and historical charging behavior. Devices with battery levels below 10% have their priority automatically increased, while devices with battery levels above 80% have their priority decreased. 4.The multi-device identification based intelligent charging priority allocation method of claim 1, wherein: The dynamic threshold adjustment strategy is optimized based on the number of charging devices and the total power load. If a new device is added, the priority score of all devices is recalculated. If a device is disconnected, the power resources occupied by that device are released and redistributed to the remaining devices. 5.The multi-device identification based intelligent charging priority allocation method of claim 1, wherein: After analyzing the power demand of the equipment, a power demand anomaly index is generated. The method for obtaining the power demand anomaly index is as follows: First, power demand data over a period of time is collected from the charging management system. From the historical power demand data, sorted in ascending order, the first quartile (Q1) is calculated: the value at the 25th percentile of the data; the third quartile (Q3) is the value at the 75th percentile of the data; IQR, or interquartile range, represents the range between Q1 and Q3. Anomaly detection boundaries are set. For detected abnormal power demand, a power demand anomaly index is calculated to measure the degree of abnormality. If the power demand falls within the normal range, the abnormality index is set to 0, and if the power demand exceeds the upper limit or the lower limit, the abnormality index is calculated according to the degree of exceeding the range, and the expression is: ; in which, is the power demand abnormality index, is the actual power demand of the current device, is the boundary value of the normal range. 6.The multi-device identification based intelligent charging priority allocation method of claim 5, wherein: The total power consumption of all current devices is monitored, and a power grid load fluctuation index is generated by analyzing the power grid load change trend in conjunction with historical data. The method for obtaining the power grid load fluctuation index is as follows: Firstly, the total power consumption data of the power grid is collected to form a time series. The time series data of the power grid load is converted to the frequency domain through Fourier transform to obtain the frequency distribution of the load power, calculate the spectral information of the load, obtain the amplitude and energy corresponding to each frequency component, distinguish the low-frequency component and the high-frequency component, calculate the proportion of the high-frequency power fluctuation in the total load, i.e. calculate the energy of each frequency component, obtain the load fluctuation amplitude in different frequency ranges, and define the power grid load fluctuation index according to the energy proportion of the high-frequency fluctuation, the expression is: ; is the energy of the high-frequency component, i.e. the total energy of the frequency component higher than the threshold in the power grid load signal, is the total energy, i.e. the total power spectrum energy of the entire load signal, is the power grid load fluctuation index. 7.The multi-device identification based intelligent charging priority allocation method of claim 6, wherein: The power demand anomaly index and grid load fluctuation index are converted into a comprehensive feature vector. This comprehensive feature vector is then used as input to a machine learning model. The machine learning model uses the prediction of the power demand fluctuation value label in the next short period of time for each set of comprehensive feature vectors as its prediction objective. The training objective is to minimize the sum of prediction errors for all power demand fluctuation value labels in the next short period of time. The machine learning model is trained until the sum of prediction errors converges, at which point the model training stops. The power demand fluctuation value in the next short period of time is determined based on the model output. The machine learning model is a multinomial regression model. 8.The multi-device identification based intelligent charging priority allocation method of claim 7, wherein: The obtained power demand fluctuation value in the near future is compared with a preset threshold. If the power demand fluctuation value in the near future is greater than or equal to the preset threshold, it indicates that the power demand fluctuation in the near future is large, and load management measures need to be taken to reduce the instantaneous power fluctuation of the power grid. If the power demand fluctuation value in the near future is less than the preset threshold, it indicates that the power demand fluctuation in the near future is small, and charging efficiency can be improved and power allocation optimized.

9. The intelligent charging priority allocation method based on multi-device identification according to claim 8, characterized in that: The adaptive power adjustment mechanism comprises calculating a power allocation deviation: ; represents a deviation between an actual power of the device and a system allocated power , is an actual charging power, i.e. a charging power currently measured by the device, is a maximum allocatable power, i.e. a maximum charging power set by the system for the device; Set the deviation threshold value ε, if , the power is normal, no need to adjust; if , the power exceeds the distribution value, need to reduce the charging power; if , the power is lower than the distribution value, need to improve the charging power.

10. An intelligent charging priority allocation system based on multi-device identification, used to implement the intelligent charging priority allocation method based on multi-device identification as described in any one of claims 1-9, characterized in that: It includes a data acquisition module, a priority score calculation module, a power prediction module, and a priority adjustment module; Data acquisition module: It uses intelligent identification technology to obtain the unique identification information of multiple devices connected to the charging system, collects the current power, battery health status, maximum supported charging power and historical charging behavior data of each device, and generates device charging demand information; Priority score calculation module: Based on the charging demand information of the devices and combined with preset weight parameters, calculate the priority score of each device, and adopt a dynamic threshold adjustment strategy to dynamically update the priority calculation model when the number of charging devices and the total power change. Power prediction module: It adopts a load prediction algorithm based on machine learning, combined with the power demand of the equipment and the power grid load, to predict the fluctuation of power demand in the near future, and adjust the maximum allocable power of each device in advance based on the prediction results. Priority adjustment module: During the charging process, the actual charging power of the device is monitored in real time and compared with the maximum allocable power. If an abnormal power allocation is detected, the charging power of different devices is dynamically adjusted or the charging order is dynamically switched based on priority through an adaptive power adjustment mechanism.

Images