AI control method and system based on charging curve optimization, and equipment medium

Abstract

The invention discloses an AI control method and system based on charging curve optimization, and an equipment medium, belongs to the technical field of battery management, and aims to solve the technical problem of how to comprehensively improve charging efficiency, battery life and power grid stability, improve power grid operation stability and promote harmonious symbiosis of an electric vehicle and a power grid. The adopted technical scheme is as follows: data acquisition and fusion: using a battery management system to collect charge state, temperature and internal resistance data of a battery, acquiring voltage and current fluctuation data through a charging pile sensor, then acquiring real-time electricity price information of a power grid by means of a network, collecting and sorting historical charging behavior data of a user, and storing the historical charging behavior data of the user; analyzing a BMS protocol by using a multi-source heterogeneous data fusion technology, cleaning unstructured user behavior data, and obtaining fused data; dynamically optimizing the charging curve of the AI drive; self-adaptive battery health management; and edge-cloud collaborative computing.

Description

Technical Field

[0001] The present invention relates to the technical field of battery management, and more specifically, to an AI control method, system, and device medium based on charging curve optimization. Background Art

[0002] With the popularization of electric vehicles, charging efficiency, battery life, and grid stability have become key issues that need to be urgently solved in the development of electric vehicles. Traditional charging technologies are no longer able to meet the rapidly developing needs of the current electric vehicle industry. In terms of charging efficiency, users expect to complete charging in a short time to reduce waiting time and improve travel convenience; in terms of battery life, frequent charging and unreasonable charging strategies will accelerate battery aging and increase the cost of battery replacement for users; in terms of grid stability, simultaneous charging of a large number of electric vehicles may impact the grid, cause grid load fluctuations, and even affect the normal operation of the grid.

[0003] Existing charging technology solutions mainly adopt a fixed charging curve mode, with the constant current - constant voltage mode being the most typical, as follows:

[0004] ① Charging process: In the initial stage of charging, the charger charges the battery with a constant current, and the battery voltage gradually rises. When the battery voltage reaches a certain threshold, the charging mode switches to constant voltage charging, and the current gradually decreases until the battery is fully charged. During the entire process, the changes in charging current and voltage follow a preset fixed curve.

[0005] ② Data utilization: Only rely on basic battery information provided by the battery management system (BMS), such as voltage, current, and state of charge (SOC), etc., for simple charging control, and hardly collect and analyze data from charging pile sensors, real - time grid electricity prices, and user historical behavior data.

[0006] ③ System interaction ability: The relationship between the charging pile and the grid is basically a one - way energy transmission. The charging pile only obtains electrical energy from the grid to charge the vehicle, lacking the ability of two - way interaction and coordinated scheduling, and unable to adjust the charging process according to factors such as grid load changes and electricity price fluctuations.

[0007] ④ Hardware and architecture characteristics: In terms of hardware design, the accuracy and layout of sensors mainly meet the basic requirements of fixed charging strategies. The communication module only ensures the transmission of basic charging status information, and is weak in low - latency communication and real - time data processing. The overall system architecture is relatively simple, mostly in a closed - loop design, making it difficult to update algorithms and expand functions, and hard to connect to new energy sources or adapt to new charging scenarios.

[0008] In summary, in the constant current-constant voltage mode, the charging process is carried out in a preset fixed manner without adjustment according to the changes in battery state, environmental conditions, and grid demand, and it cannot dynamically adapt to battery state, environmental conditions, and grid demand. Since the state of the battery (such as state of charge SOC, temperature, internal resistance, etc.) changes during different usage stages, the fixed charging curve cannot adjust the charging power in real time according to these changes, resulting in rapid battery loss. For example, when the battery temperature is too high, if charging is still carried out at a large current according to the fixed curve, it will accelerate the chemical reaction inside the battery and shorten the battery life. At the same time, this "one-size-fits-all" charging strategy will cause a large impact on the grid when a large number of electric vehicles charge simultaneously during the peak grid load period, increasing the pressure on grid expansion.

[0009] In terms of fast charging technology, the existing technology usually sacrifices battery life. Although high-rate charging can replenish the battery power in a short time, it will cause problems such as lithium precipitation, accelerating battery aging. For example, during fast charging, the movement speed of lithium ions inside the battery is too fast, which is easy to form lithium dendrites on the surface of the negative electrode, piercing the diaphragm, causing safety problems such as battery short circuit, and at the same time reducing the cycle life of the battery. In addition, most of the existing charging piles are for one-way energy transmission, lacking the ability of coordinated scheduling with the grid, unable to carry out charging scheduling elastically according to the peak and valley of the grid load and user demand, unable to effectively utilize grid resources, and not conducive to the stable operation of the grid.

[0010] Therefore, the existing charging technologies have the following defects:

[0011] ① The problem of poor adaptability of traditional charging strategies: The existing technology relies on fixed charging curves (such as the constant current-constant voltage mode) and cannot dynamically adjust the charging strategy according to battery state (SOC, temperature, internal resistance, etc.), environmental conditions, and grid demand. This makes it difficult to achieve the best charging for the battery under different working conditions, resulting in increased battery loss, reduced charging efficiency, and a large impact on the grid. The present invention uses AI technology to analyze multi-source data in real time, dynamically generate the optimal charging power curve, realize the precise control of the charging strategy, take into account efficiency, safety, and economy, and solve a series of problems caused by the "one-size-fits-all" charging strategy.

[0012] ② The contradiction between battery life and charging speed: Although fast charging technology can meet the user's demand for rapid energy replenishment, high-rate charging is likely to cause problems such as lithium precipitation, accelerating battery aging and significantly shortening the service life of the battery. The present invention uses AI to predict the microscopic aging mechanism of the battery (such as SEI film growth, mechanical stress distribution), and by dynamically adjusting the current / voltage curve, while ensuring the charging speed, effectively inhibits battery aging, achieves the goal of "fast charging without damaging the battery", extends the overall service life of the battery, and reduces the user's battery replacement cost.

[0013] ③ Lack of grid interaction and coordinated dispatching capabilities: Traditional charging piles are mostly one-way energy transmission, lacking the ability to coordinate with the grid, and are unable to cope with the load pressure on the grid caused by the centralized charging of a large number of electric vehicles. During the peak load period of the grid, many charging piles draw power at the same time, which can easily cause grid voltage fluctuations, power imbalance and other problems, increasing the cost of grid expansion.

[0014] Therefore, how to comprehensively improve charging efficiency, battery life and grid stability, improve grid operation stability, and promote harmonious coexistence of electric vehicles and grids is a technical problem that needs to be solved urgently. Summary of the invention

[0015] The technical task of the present invention is to provide an AI control method, system, and device medium based on charging curve optimization to solve the problem of how to comprehensively improve charging efficiency, battery life, and grid stability, improve grid operation stability, and promote the harmonious coexistence of electric vehicles and grids.

[0016] The technical task of the present invention is achieved in the following way: an AI control method based on charging curve optimization, the method is as follows:

[0017] Data collection and fusion: The battery management system (BMS) is used to collect the battery's state of charge (SOC), temperature, and internal resistance data. At the same time, the voltage and current fluctuation data are obtained through the charging pile sensor. The real-time electricity price information of the power grid is obtained through the network. The user's historical charging behavior data is collected and sorted. Then, multi-source heterogeneous data fusion technology is used to parse the BMS protocol, clean the unstructured user behavior data, and obtain the fused data;

[0018] AI-driven dynamic optimization of charging curves: Machine learning is used to analyze the fused data in real time, and a charging power curve that satisfies multiple objectives is dynamically generated based on the analysis results, balancing the shortest charging time, lowest battery power consumption, lowest grid load fluctuation, and lowest user cost during the charging process.

[0019] Adaptive battery health management: Predict battery aging trends through AI algorithms and develop differentiated charging strategies for different battery types;

[0020] Edge-cloud collaborative computing: A lightweight model is used locally to achieve low-latency control of the charging process through edge computing to ensure real-time adjustment of charging parameters. At the same time, historical data is uploaded to the cloud, and the model is trained using cloud big data to continuously iterate and optimize to continuously improve performance.

[0021] Preferably, the method also configures a high-precision sensor array, optimizes the layout of temperature and current sampling points, accurately collects data, and designs a low-latency communication module to ensure that AI commands are transmitted to the charging device quickly and accurately to achieve real-time control.

[0022] Preferably, the differential charging strategy is specifically as follows: when the battery is approaching the high SOC range, if it is predicted that fast charging will accelerate aging, the charging current and voltage are adjusted to avoid fast charging in the high SOC range and extend the battery life.

[0023] Preferably, the data fusion and acquisition are specifically as follows:

[0024] Battery information acquisition: Connect to the battery management system (BMS) to obtain key parameters such as the state of charge (SOC), voltage, current, temperature, and internal resistance of the battery in real time;

[0025] Charging pile parameter acquisition: Acquire the voltage, current, and power data output by the charging pile, and monitor the working state of the charging pile at the same time; among them, the working state of the charging pile includes whether it is operating normally and whether there are any faults;

[0026] Environmental information acquisition: Obtain the temperature and humidity environmental parameters of the charging environment through temperature sensors and humidity sensors;

[0027] Grid information acquisition: Connect to the grid through a communication interface to obtain the real-time electricity price and load conditions of the grid;

[0028] Data cleaning and preprocessing: Perform denoising, outlier processing, and normalization operations on the collected raw data to ensure the accuracy and usability of the data;

[0029] Feature extraction: Extract key features related to charging from the processed data; among them, the key features related to charging include the change trend of the charging and discharging rate of the battery and the influence coefficient of the environmental temperature on the battery performance.

[0030] More preferably, the AI-driven dynamic optimization of the charging curve is specifically as follows:

[0031] Build a charging strategy decision model: Use the deep reinforcement learning algorithm to build a charging strategy decision model, and the built charging strategy decision model is trained through historical charging data and real-time collected data; during the training process, the built charging strategy decision model continuously tries different charging strategies and optimizes its own strategy according to the reward mechanism (such as improved charging efficiency, extended battery life, grid load balancing, etc.);

[0032] Strategy generation: Generate the optimal charging strategy for the current charging state according to the real-time data and the trained charging strategy decision model; among them, the optimal charging strategy for the current charging state includes the adjustment scheme of the charging current and voltage and the planning of the charging time;

[0033] Signal conversion and transmission: Convert the charging strategy into a control signal and transmit it to the controller of the charging pile through a communication interface;

[0034] Charging pile control: After receiving the control signal, the charging pile controller adjusts the charging parameters in real time according to the instructions to achieve precise control of the charging process.

[0035] An AI control system based on charging curve optimization. The system includes a data acquisition module, a data processing and analysis module, an AI decision-making module, and a charging control module. The data acquisition module transmits the collected data to the data processing and analysis module. After processing and analysis, the data is input into the AI decision-making module. The charging strategy generated by the AI decision-making module acts on the charging pile through the charging control module to achieve the charging control of the electric vehicle, forming a complete closed-loop control system.

[0036] Among them, the data acquisition module is used to collect the state of charge (SOC), temperature, and internal resistance data of the battery by using the battery management system (BMS). At the same time, it obtains voltage and current fluctuation data through the charging pile sensor, obtains the real-time electricity price information of the power grid through the network, and collects and collates the historical charging behavior data of users.

[0037] The data processing and analysis module is used to perform denoising, outlier processing, and normalization operations on the collected raw data, and extract key features related to charging from the processed data.

[0038] The AI decision-making module is used to perform real-time analysis on the fused data by using machine learning, and dynamically generate a charging power curve that meets multi-objective optimization, balancing the shortest charging time, the lowest battery power consumption, the lowest power grid load fluctuation, and the lowest user cost during the charging process.

[0039] The charging control module is used to adopt a lightweight model locally to achieve low-latency control of the charging process through edge computing, ensuring real-time adjustment of charging parameters. At the same time, it uploads the historical data to the cloud, trains the model by using cloud big data, continuously iterates and optimizes, and continuously improves the performance.

[0040] Preferably, the data acquisition module includes:

[0041] The battery information acquisition sub-module is used to connect to the battery management system (BMS) to obtain the key parameters of the state of charge (SOC), voltage, current, temperature, and internal resistance of the battery in real time.

[0042] The charging pile parameter acquisition sub-module is used to collect the voltage, current, and power data output by the charging pile, and monitor the working state of the charging pile. Among them, the working state of the charging pile includes whether it is operating normally and whether there is a fault.

[0043] The environmental information acquisition sub-module is used to obtain the temperature and humidity environmental parameters of the charging environment through temperature sensors and humidity sensors.

[0044] The power grid information acquisition sub-module is used to connect to the power grid through a communication interface to obtain the real-time electricity price and load conditions of the power grid.

[0045] Preferably, the AI decision-making module includes:

[0046] The model construction sub-module is used to construct a charging strategy decision-making model using a deep reinforcement learning algorithm. The charging strategy decision-making model is trained through historical charging data and real-time collected data. During the training process, the charging strategy decision continuously tries different charging strategies and optimizes its own strategy according to the reward mechanism (such as improved charging efficiency, extended battery life, power grid load balancing, etc.);

[0047] The strategy generation sub-module is used to generate the optimal charging strategy for the current charging state according to the real-time data and the trained charging strategy decision-making model; among them, the optimal charging strategy for the current charging state includes the adjustment scheme of charging current and voltage and the planning of charging time;

[0048] The charging control module includes:

[0049] The signal conversion and transmission sub-module is used to convert the charging strategy generated by the AI decision-making module into a control signal and transmit it to the controller of the charging pile through a communication interface;

[0050] The charging pile control sub-module is used to, after receiving the control signal, the charging pile controller adjusts the charging parameters in real time according to the instruction to achieve precise control of the charging process.

[0051] An electronic device includes: a memory and at least one processor;

[0052] Wherein, a computer program is stored on the memory;

[0053] The at least one processor executes the computer program stored in the memory, so that the at least one processor executes the AI control method based on charging curve optimization as described above.

[0054] A computer-readable storage medium stores a computer program, and the computer program can be executed by a processor to implement the AI control method based on charging curve optimization as described above.

[0055] The AI control method, system, device, and medium based on charging curve optimization of the present invention have the following advantages:

[0056] (1) The present invention predicts the peak and valley of the power grid load through AI, combines the user demand elasticity, and realizes demand response, such as preferentially charging during the low valley period, effectively reducing the pressure of power grid expansion, improving the operation stability of the power grid, and promoting the harmonious coexistence of electric vehicles and the power grid. It can take into account charging efficiency, battery life, and power grid stability;

[0057] (ii) The present invention designs a charging system that supports multiple charging standards (such as CCS, CHAdeMO, GB / T), so that it can adapt to different vehicle models and charging scenarios (public charging piles, private charging piles, supercharging stations), and adopts a modular architecture to facilitate the subsequent access to new energy sources (such as integrated photovoltaic storage and charging), or upgrade new algorithms to enhance system functions;

[0058] (III) Unlike the prior art which only relies on basic data from the battery management system (BMS), the present invention not only collects basic data such as the battery's state of charge (SOC), voltage, current, temperature, internal resistance, etc., but also widely collects multi-source data such as charging pile operating parameters, charging environment information (temperature, humidity, etc.), and real-time power grid electricity prices and load conditions. This comprehensive data collection method provides a richer and more comprehensive information basis for subsequent AI analysis;

[0059] (IV) The present invention uses advanced data fusion technology to efficiently integrate heterogeneous data from different data sources; at the same time, it uses complex data mining and feature extraction algorithms to extract key features that can reflect battery health status, charging efficiency, grid adaptability, etc. from the fused data. These features are not paid attention to and used by traditional charging technologies, providing accurate input for AI decision-making;

[0060] (V) The present invention uses a deep reinforcement learning algorithm to construct a charging strategy decision model. This model is different from the traditional fixed charging strategy and can optimize its own decision-making by constantly interacting with the environment (i.e., adjusting the charging strategy according to real-time data and obtaining reward feedback according to the charging effect); at the same time, deep learning is combined to conduct in-depth analysis of a large amount of historical charging data, and potential laws and patterns in the data are mined, thereby realizing dynamic optimization of the charging curve;

[0061] (VI) When generating the charging curve, the AI ​​model of the present invention comprehensively considers multiple objectives, including the shortest charging time, the lowest battery loss, the lowest grid load fluctuation, and the lowest user cost. Traditional charging technologies can often only focus on one or two objectives and cannot achieve balance and coordinated optimization among multiple objectives. The present invention can find the optimal charging strategy under different charging scenarios and conditions through the intelligent decision-making of the AI ​​algorithm, and achieve the best balance among multiple objectives.

[0062] (VII) The present invention uses AI technology to accurately predict the aging trend of the battery. By analyzing the data changes of the battery during different charge and discharge cycles and combining the physical and chemical model of the battery, the microscopic aging mechanism inside the battery, such as the growth of the solid electrolyte interface (SEI) film and the distribution of mechanical stress, is predicted. This prediction capability is not available in traditional charging technology, and measures can be taken in advance to delay battery aging.

[0063] (8) The present invention customizes personalized charging strategies for each battery according to the characteristics of different battery types (such as lithium iron phosphate, ternary lithium, etc.) and the real-time health status of the battery. For example, when the battery is approaching the high SOC range, if it is predicted that fast charging will accelerate aging, the charging current and voltage are automatically adjusted to avoid fast charging in the high SOC range, thereby extending the overall service life of the battery.

[0064] (9) The present invention deploys edge computing devices locally and runs lightweight AI models to achieve low-latency real-time control of the charging process, and can quickly respond to real-time changes during the charging process, such as sudden changes in battery status and fluctuations in grid voltage, and timely adjust charging parameters to ensure the stability and safety of the charging process.

[0065] (10) Upload a large amount of historical charging data to the cloud, utilize the powerful computing resources of the cloud for big data analysis and in-depth mining, and through learning from a large amount of data, continuously iterate and optimize the AI model to improve the accuracy and adaptability of the model, so that the system can continuously adapt to new charging scenarios and requirements, which cannot be achieved by traditional centralized or local computing architectures. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] The present invention will be further described below with reference to the accompanying drawings.

[0067] Attached Figure 1 is a schematic structural diagram of an AI control system optimized based on a charging curve. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0068] The AI control method, system, and device medium optimized based on the charging curve of the present invention will be described in detail below with reference to the accompanying drawings of the specification and specific embodiments.

[0069] Embodiment 1:

[0070] This embodiment provides an AI control method optimized based on a charging curve, and the method is as follows:

[0071] S1. Data collection and fusion: Use the battery management system (BMS) to collect the state of charge (SOC), temperature, and internal resistance data of the battery. At the same time, obtain voltage and current fluctuation data through the charging pile sensor, obtain the real-time grid electricity price information through the network, collect and sort out the user's historical charging behavior data, and then use the multi-source heterogeneous data fusion technology to analyze the BMS protocol and clean the unstructured user behavior data to obtain the fused data.

[0072] S2. AI-driven Dynamic Optimization of Charging Curve: Use machine learning to perform real-time analysis on the fused data, and dynamically generate a charging power curve that meets multi-objective optimization, balancing the shortest charging time, the lowest battery power consumption, the lowest grid load fluctuation, and the lowest user cost during the charging process; for example, when the battery temperature is detected to be high, reduce the charging power to reduce battery loss; if it is in the low grid load period and the user has no urgent charging demand, appropriately increase the charging power to speed up the charging speed; machine learning algorithms include reinforcement learning and deep learning.

[0073] S3. Adaptive Battery Health Management: Use AI algorithms to predict the battery aging trend, and formulate differentiated charging strategies for different battery types (such as lithium iron phosphate, ternary lithium, etc.); when the battery is approaching the high SOC range, if it is predicted that fast charging will accelerate aging, adjust the charging current / voltage to avoid fast charging in the high SOC range and extend the battery life.

[0074] S4. Edge-Cloud Collaborative Computing: Adopt a lightweight model locally, and achieve low-latency control of the charging process through edge computing to ensure real-time adjustment of charging parameters; at the same time, upload historical data to the cloud, use cloud big data to train the model, continuously iterate and optimize, and continuously improve performance.

[0075] This embodiment also configures a high-precision sensor array, optimizes the layout of temperature and current sampling points, accurately collects data, and designs a low-latency communication module to ensure that AI instructions are quickly and accurately transmitted to the charging device to achieve real-time control.

[0076] The differentiated charging strategy in step S3 of this embodiment is specifically: when the battery is approaching the high SOC range, if it is predicted that fast charging will accelerate aging, adjust the charging current and voltage to avoid fast charging in the high SOC range and extend the battery life.

[0077] The data fusion and collection in step S1 of this embodiment are specifically as follows:

[0078] S101. Battery Information Collection: Connect to the battery management system (BMS) to obtain key parameters such as the state of charge (SOC), voltage, current, temperature, and internal resistance of the battery in real time; for example, during the driving of an electric vehicle, collect battery SOC data every 500 milliseconds to provide basic battery state information for subsequent charging strategy formulation.

[0079] S102. Charging Pile Parameter Collection: Collect the voltage, current, and power data output by the charging pile, and monitor the working state of the charging pile at the same time; among them, the working state of the charging pile includes whether it is running normally and whether there is a fault; for example, collect the charging pile output current data every 1 second to judge whether the charging process is stable.

[0080] S103. Environmental information collection: Obtain the temperature and humidity environmental parameters of the charging environment through temperature sensors and humidity sensors; in the outdoor charging scenario, the environmental temperature changes can be sensed in real time to adjust the charging strategy in high or low temperature environments;

[0081] S104. Grid information collection: Connect to the grid through a communication interface to obtain the real-time electricity price and load conditions of the grid; during the peak load period of the grid, the electricity price is high and the power supply pressure is large, and the charging plan can be adjusted accordingly;

[0082] S105. Data cleaning and preprocessing: Perform denoising, outlier processing, and normalization operations on the collected raw data to ensure the accuracy and availability of the data; for example, when detecting abnormal jumps in battery voltage data, remove noise interference through a filtering algorithm;

[0083] S106. Feature extraction: Extract key features related to charging from the processed data; among them, the key features related to charging include the change trend of the charging and discharging rate of the battery and the influence coefficient of the environmental temperature on the battery performance.

[0084] The AI-driven dynamic optimization of the charging curve in step S2 of this embodiment is specifically as follows:

[0085] S201. Build a charging strategy decision model: Use a deep reinforcement learning algorithm to build a charging strategy decision model, and the charging strategy decision model is trained through historical charging data and real-time collected data; during the training process, the charging strategy decision model continuously tries different charging strategies and optimizes its own strategy according to the reward mechanism (such as improved charging efficiency, extended battery life, grid load balancing, etc.); for example, when the model predicts that increasing the charging power during the current grid load valley period can improve the charging efficiency without affecting the battery life, it will adjust the charging strategy to increase the charging power;

[0086] S202. Strategy generation: Generate the optimal charging strategy for the current charging state according to the real-time data and the trained charging strategy decision model; among them, the optimal charging strategy for the current charging state includes the adjustment scheme of the charging current and voltage and the planning of the charging time;

[0087] S203. Signal conversion and transmission: Convert the charging strategy into a control signal and transmit it to the controller of the charging pile through a communication interface; for example, convert the instruction to adjust the charging current into a digital signal and send it to the charging pile;

[0088] S204. Charging pile control: After receiving the control signal, the charging pile controller adjusts the charging parameters in real time according to the instruction to achieve precise control of the charging process; for example, when the control signal requires reducing the charging current, the charging pile controller adjusts the circuit components to reduce the output current.

[0089] This embodiment also adopts the following technologies:

[0090] ① Multi-source data fusion and processing method: including the scope, method and frequency of data collection, as well as the algorithms and processes of data fusion, the specific methods of feature extraction and the key features extracted.

[0091] ② AI charging strategy decision-making model: covering the architecture of the model, the combined way of the adopted reinforcement learning and deep learning algorithms, the methods and processes of model training, and the mechanisms and processes of the model generating charging strategies.

[0092] ③ Adaptive battery health management technology: involving the algorithms and models for battery aging prediction, the formulation principles and methods of personalized charging strategies, and the mechanism for dynamically adjusting charging strategies according to the battery health status.

[0093] ④ Edge-cloud collaborative computing architecture: including the function division of edge computing devices and cloud servers, data transmission protocols and interaction methods, the design and operation mechanisms of lightweight models at the edge, and the processes and methods of cloud big data analysis and model iteration.

[0094] This embodiment can also be applied from fixed network technology to mobile network, specifically as follows:

[0095] ① In terms of data collection: In the fixed network, the operation parameters of network devices (such as routers, switches, etc.) are mainly collected, such as port traffic, link status, etc. In the mobile network, the data collection objects become base stations and mobile terminals. It is necessary to newly collect data such as base station signal strength, interference situation, and the location, battery power, signal reception quality, etc. of mobile terminals. For example, use the GPS module of the mobile terminal to obtain location information, and collect battery power information through built-in sensors. At the same time, it is necessary to optimize the data collection frequency and method to adapt to the characteristics of rapid movement of devices and dynamic signal changes in the mobile network.

[0096] ② AI decision-making module: In the fixed network, AI is mainly used for network traffic optimization, fault diagnosis, etc., and its decision-making basis is relatively stable. In the mobile network, AI decision-making needs to consider more dynamic factors. For example, since the location of the mobile terminal is constantly changing, AI needs to dynamically decide which base station the terminal accesses and how to allocate communication resources in real time according to the terminal location and the load situation of surrounding base stations. When dealing with communication congestion, use AI to analyze the user distribution and service requirements in different regions, and intelligently adjust parameters such as the transmission power and channel allocation of base stations to ensure the communication quality of key services is preferentially guaranteed in high-traffic regions.

[0097] ③ Control Execution Module: The control in the fixed network mainly focuses on the configuration adjustment of network devices, such as modifying the routing table of routers, port settings of switches, etc. In the mobile network, control execution needs to closely cooperate with base stations and mobile terminals. For example, when the AI decision-making module determines that a base station handover is required, the control execution module sends handover instructions to the mobile terminal and the target base station, coordinating both sides to complete signal handover and resource reallocation. At the same time, it is necessary to ensure the accuracy and timeliness of control instructions during wireless transmission, avoiding control failures caused by signal interference and other problems.

[0098] In addition to the mobile network, this embodiment also has application prospects in other fields. For example, in the field of smart grids, by collecting operation data of power equipment, user electricity consumption habit data, etc., AI can be used to optimize power dispatching strategies to achieve efficient allocation of power resources and stable operation of the power grid; in the field of smart homes, by collecting operation data of various smart home appliances and user living habit data, AI can intelligently control the operation of home appliances based on these data to achieve an energy-saving and comfortable home environment.

[0099] Embodiment 2:

[0100] As shown Figure 1 in the figure, this embodiment provides an AI control system based on charging curve optimization. The system includes a data acquisition module, a data processing and analysis module, an AI decision-making module, and a charging control module. The data acquisition module transmits the collected data to the data processing and analysis module. The data after processing and analysis is input into the AI decision-making module. The charging strategy generated by the AI decision-making module acts on the charging pile through the charging control module to achieve the charging control of electric vehicles, forming a complete closed-loop control system;

[0101] Among them, the data acquisition module is used to collect the state of charge (SOC), temperature, and internal resistance data of the battery using the battery management system (BMS). At the same time, it obtains voltage and current fluctuation data through the charging pile sensor, and obtains the real-time electricity price information of the power grid through the network, and collects and organizes the historical charging behavior data of users;

[0102] The data processing and analysis module is used to perform denoising, outlier processing, and normalization operations on the collected original data, and extract key charging-related features from the processed data;

[0103] The AI decision-making module is used to perform real-time analysis on the fused data using machine learning, and dynamically generate a charging power curve that meets multi-objective optimization, balancing the shortest charging time, the lowest battery power consumption, the lowest power grid load fluctuation, and the lowest user cost during the charging process;

[0104] The charging control module is used to adopt a lightweight model locally and achieve low-latency control of the charging process through edge computing to ensure real-time adjustment of charging parameters. At the same time, it uploads historical data to the cloud, trains the model using cloud big data, continuously iterates and optimizes, and continuously improves performance.

[0105] The data acquisition module in this embodiment includes:

[0106] The battery information acquisition sub-module is used to connect to the battery management system (BMS) to obtain key parameters such as the state of charge (SOC), voltage, current, temperature, and internal resistance of the battery in real time. For example, during the driving of an electric vehicle, the battery SOC data is collected every 500 milliseconds to provide basic battery state information for subsequent charging strategy formulation.

[0107] The charging pile parameter acquisition sub-module is used to collect voltage, current, and power data output by the charging pile, and at the same time monitor the working state of the charging pile. Among them, the working state of the charging pile includes whether it is operating normally and whether there are faults. For example, the charging pile output current data is collected every 1 second to judge whether the charging process is stable.

[0108] The environmental information acquisition sub-module is used to obtain temperature and humidity environmental parameters of the charging environment through temperature sensors and humidity sensors. In an outdoor charging scenario, it can perceive changes in the environmental temperature in real time to adjust the charging strategy in high-temperature or low-temperature environments.

[0109] The power grid information acquisition sub-module is used to connect to the power grid through a communication interface to obtain the real-time electricity price and load conditions of the power grid. When the power grid load is at a peak, the electricity price is high and the power supply pressure is large, and the system can adjust the charging plan accordingly.

[0110] The AI decision-making module in this embodiment includes:

[0111] The model construction sub-module is used to construct a charging strategy decision-making model using the deep reinforcement learning algorithm. The charging strategy decision-making model is trained with historical charging data and real-time collected data. During the training process, the charging strategy decision continuously tries different charging strategies and optimizes its own strategy according to the reward mechanism (such as improved charging efficiency, extended battery life, power grid load balancing, etc.).

[0112] The strategy generation sub-module is used to generate the optimal charging strategy for the current charging state according to the real-time data and the trained charging strategy decision-making model. Among them, the optimal charging strategy for the current charging state includes the adjustment scheme of charging current and voltage and the planning of charging time.

[0113] The charging control module in this embodiment includes:

[0114] The signal conversion and transmission sub-module is used to convert the charging strategy generated by the AI decision-making module into a control signal and transmit it to the controller of the charging pile through the communication interface; for example, converting the instruction to adjust the charging current into a digital signal and sending it to the charging pile.

[0115] The charging pile control sub-module is used to receive the control signal, and then the charging pile controller adjusts the charging parameters in real time according to the instruction to achieve precise control of the charging process; for example, when the control signal requires reducing the charging current, the charging pile controller adjusts the circuit elements to reduce the output current.

[0116] The working process of this system is as follows:

[0117] (1) Data acquisition: According to the functions of the above data acquisition module, periodically acquire relevant data of the battery, charging pile, environment, and power grid; for example, after the electric vehicle is connected to the charging pile, immediately start the data acquisition process, and acquire data at a high frequency with a short time interval (such as in seconds) to ensure timely acquisition of the initial charging state information.

[0118] (2) Data processing and analysis step: Perform cleaning and feature extraction operations on the acquired raw data in sequence; in the data cleaning stage, use statistical methods to identify and remove abnormal data points; in the feature extraction stage, adopt data mining algorithms to extract feature values reflecting aspects such as the battery health state and charging efficiency.

[0119] (3) AI decision-making: Input the processed and analyzed data into the trained machine learning model, and the model generates the optimal charging strategy based on the current data and the learned knowledge; for example, in the initial stage of model training, explore various possible charging strategies and their effects by simulating the charging process in a large number of different scenarios. As the training progresses, the model gradually learns to select the best strategy under different conditions.

[0120] (4) Charging control: After the charging control module receives the charging strategy generated by the AI decision-making module, it converts it into a specific control signal and sends it to the charging pile. The charging pile adjusts the charging parameters in real time according to the control signal to complete the control of the entire charging process; during the charging process, continuously monitor the data and return it to the AI decision-making module for dynamic adjustment to ensure that the charging process is always in the optimal state.

[0121] Embodiment 3:

[0122] This embodiment also provides an electronic device, including: a memory and a processor;

[0123] Among them, the memory stores computer execution instructions;

[0124] The processor executes the computer-executable instructions stored in the memory, causing the processor to execute the AI control method optimized based on the charging curve in any embodiment of the present invention.

[0125] The processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), off-the-shelf programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0126] The memory can be used to store computer programs and / or modules. The processor realizes various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory, and by calling the data stored in the memory. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc.; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory may also include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one magnetic disk storage period, flash device, or other volatile solid-state storage devices.

[0127] Embodiment 4:

[0128] This embodiment also provides a computer-readable storage medium, which stores multiple instructions. The instructions are loaded by the processor, causing the processor to execute the AI control method optimized based on the charging curve in any embodiment of the present invention. Specifically, a system or device equipped with a storage medium can be provided. On the storage medium, software program codes for implementing the functions of any one of the above embodiments are stored, and the computer (or CPU or MPU) of the system or device reads and executes the program codes stored in the storage medium.

[0129] In this case, the program code read from the storage medium itself can implement the functions of any one of the above embodiments. Therefore, the program code and the storage medium storing the program code constitute a part of the present invention.

[0130] Embodiments of the storage medium for providing program codes include floppy disks, hard disks, magneto-optical disks, optical disks (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RYM, DVD-RW, DVD+RW), magnetic tapes, non-volatile memory cards, and ROMs. Optionally, the program code can be downloaded from a server computer via a communication network.

[0131] In addition, it should be clear that not only can the above-described functions of any one of the embodiments be achieved by executing program codes read by a computer, but also by means of instructions based on the program codes to cause an operating system or the like operating on the computer to perform part or all of the actual operations.

[0132] In addition, it can be understood that the program codes read from the storage medium are written into a memory provided in an expansion board inserted into the computer or into a memory provided in an expansion unit connected to the computer, and then, based on the instructions of the program codes, a CPU or the like installed on the expansion board or the expansion unit is caused to perform part and all of the actual operations, thereby implementing the functions of any one of the above-described embodiments.

[0133] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An AI control method based on charging curve optimization, characterized in that: The method is as follows: Data collection and fusion: The battery management system is used to collect the battery's state of charge, temperature, and internal resistance data. At the same time, the voltage and current fluctuation data are obtained through the charging pile sensor. The real-time electricity price information of the power grid is obtained through the network. The user's historical charging behavior data is collected and sorted. Then, multi-source heterogeneous data fusion technology is used to analyze the BMS protocol, clean the unstructured user behavior data, and obtain the fused data; AI-driven dynamic optimization of charging curves: Machine learning is used to analyze the fused data in real time, and a charging power curve that satisfies multiple objectives is dynamically generated based on the analysis results, balancing the shortest charging time, lowest battery power consumption, lowest grid load fluctuation, and lowest user cost during the charging process. Adaptive battery health management: Predict battery aging trends through AI algorithms and develop differentiated charging strategies for different battery types; Edge-cloud collaborative computing: A lightweight model is used locally to achieve low-latency control of the charging process through edge computing to ensure real-time adjustment of charging parameters. At the same time, historical data is uploaded to the cloud, and the model is trained using cloud big data to continuously iterate and optimize to continuously improve performance.

2. The AI ​​control method based on charging curve optimization according to claim 1 is characterized in that: This method also configures a high-precision sensor array, optimizes the layout of temperature and current sampling points, accurately collects data, and designs a low-latency communication module to ensure that AI commands are transmitted to the charging device quickly and accurately to achieve real-time control.

3. The AI ​​control method based on charging curve optimization according to claim 1 is characterized in that: The differentiated charging strategy is as follows: when the battery approaches a high SOC range, if it is predicted that fast charging will accelerate aging, the charging current and voltage are adjusted to avoid fast charging in the high SOC range and extend the battery life.

4. The AI ​​control method based on charging curve optimization according to claim 1, characterized in that: The details of data fusion and collection are as follows: Battery information collection: Connect to the battery management system to obtain key parameters of the battery's state of charge, voltage, current, temperature and internal resistance in real time; Charging pile parameter collection: Collect the voltage, current and power data output by the charging pile, and monitor the working status of the charging pile; the working status of the charging pile includes whether it is operating normally and whether there is a fault; Environmental information collection: obtain the temperature and humidity parameters of the charging environment through temperature sensors and humidity sensors; Grid information collection: connect to the grid through the communication interface to obtain the real-time electricity price and load conditions of the grid; Data cleaning and preprocessing: denoising, outlier processing and normalization operations are performed on the collected raw data to ensure the accuracy and availability of the data; Feature extraction: Extract key features related to charging from the processed data; the key features related to charging include the changing trend of the battery's charge and discharge rate and the influence coefficient of ambient temperature on battery performance.

5. The AI ​​control method based on charging curve optimization according to any one of claims 1 to 4, characterized in that: The AI-driven dynamic optimization of charging curve is as follows: Constructing a charging strategy decision model: A deep reinforcement learning algorithm is used to construct a charging strategy decision model. The charging strategy decision model is trained through historical charging data and real-time collected data. During the training process, the charging strategy decision model continuously tries different charging strategies and optimizes its own strategy based on the reward mechanism. Strategy generation: Generate the optimal charging strategy for the current charging state based on real-time data and the trained charging strategy decision model; the optimal charging strategy for the current charging state includes the adjustment plan of charging current and voltage and the planning of charging time; Signal conversion and transmission: convert the charging strategy into a control signal and transmit it to the controller of the charging pile through the communication interface; Charging pile control: After receiving the control signal, the charging pile controller adjusts the charging parameters in real time according to the instructions to achieve precise control of the charging process.

6. An AI control system based on charging curve optimization, characterized in that: The system includes a data acquisition module, a data processing and analysis module, an AI decision module and a charging control module. The data acquisition module transmits the collected data to the data processing and analysis module, and the processed and analyzed data is input into the AI ​​decision module. The charging strategy generated by the AI ​​decision module acts on the charging pile through the charging control module to realize the charging control of the electric vehicle, forming a complete closed-loop control system. The data acquisition module is used to collect the battery state of charge, temperature and internal resistance data using the battery management system, and to obtain voltage and current fluctuation data through the charging pile sensor, and then obtain the real-time electricity price information of the power grid through the network, and collect and organize the user's historical charging behavior data; Data processing and analysis module, used to perform denoising, outlier processing and normalization operations on the collected raw data, and extract key features related to charging from the processed data; The AI ​​decision-making module uses machine learning to perform real-time analysis on the fused data and dynamically generates a charging power curve that meets multiple objective optimizations based on the analysis results, balancing the shortest charging time, lowest battery power consumption, lowest grid load fluctuation, and lowest user cost during the charging process; The charging control module is used to adopt a lightweight model locally and achieve low-latency control of the charging process through edge computing to ensure real-time adjustment of charging parameters. At the same time, historical data is uploaded to the cloud, and the cloud big data training model is used to continuously iterate and optimize to continuously improve performance.

7. The AI ​​control system based on charging curve optimization according to claim 6, characterized in that: The data acquisition module includes: The battery information acquisition submodule is used to connect to the battery management system to obtain key parameters of the battery's state of charge, voltage, current, temperature, and internal resistance in real time; The charging pile parameter collection submodule is used to collect the voltage, current and power data output by the charging pile, and monitor the working status of the charging pile; the working status of the charging pile includes whether it is operating normally and whether there is a fault; The environmental information acquisition submodule is used to obtain the temperature and humidity environmental parameters of the charging environment through the temperature sensor and the humidity sensor; The power grid information collection submodule is used to connect to the power grid through the communication interface to obtain the real-time electricity price and load conditions of the power grid.

8. The AI ​​control system based on charging curve optimization according to claim 6 or 7, characterized in that: AI decision-making modules include: The model building submodule is used to build a charging strategy decision model using a deep reinforcement learning algorithm. The charging strategy decision model is trained using historical charging data and real-time collected data. During the training process, the charging strategy decision continuously tries different charging strategies and optimizes its own strategy based on the reward mechanism. The strategy generation submodule is used to generate the optimal charging strategy for the current charging state based on real-time data and the trained charging strategy decision model; wherein the optimal charging strategy for the current charging state includes the adjustment scheme of the charging current and voltage and the planning of the charging time; The charging control module includes: The signal conversion and transmission submodule is used to convert the charging strategy generated by the AI ​​decision module into a control signal and transmit it to the controller of the charging pile through the communication interface; The charging pile control submodule is used to receive the control signal, and the charging pile controller adjusts the charging parameters in real time according to the instructions to achieve precise control of the charging process.

9. An electronic device, characterized in that: include: memory and at least one processor; Wherein, the memory stores a computer program; The at least one processor executes the computer program stored in the memory, so that the at least one processor performs the AI ​​control method based on charging curve optimization according to any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which can be executed by a processor to implement the AI ​​control method based on charging curve optimization as described in any one of claims 1 to 5.

Images