Intelligent charging control method, device and system

Abstract

The invention relates to the technical field of intelligent charging, in particular to an intelligent charging control method, device and system, and the method comprises the following specific steps: (1) collecting and fusing multi-dimensional data; (2) dynamically generating an adaptive charging window; (3) multi-objective optimization strategy decision making; and (4) executing distributed cooperative control. According to the intelligent charging control method, device and system, through technical innovation and architecture optimization, the core problems of traditional charging control in the aspects of precision, efficiency, safety, collaboration and the like are solved, a closed-loop system of data perception, intelligent decision making, collaborative execution and feedback optimization is formed, and the intelligent charging control method, device and system have the advantages of being high in practicability and high in practicability. A prospective and practical solution is provided for a new generation of charging technology, and technical progress and industrial application in the field of intelligent charging are promoted.

Description

Technical Field

[0001] This invention relates to the field of intelligent charging technology, and in particular to an intelligent charging control method, device and system. Background Technology

[0002] With the continuous development of new energy technologies, intelligent charging has become an indispensable part of electric vehicles.

[0003] Traditional charging technology has at least the following drawbacks: 1. Limitations of traditional charging technology Single-objective control strategies (such as constant current and constant voltage) cannot simultaneously consider charging speed, battery life, and grid adaptability, leading to accelerated battery aging (cycle life < 800 times) and high peak and valley load pressure on the grid. Low accuracy in battery state estimation (SOC error ≥2%, SOH error ≥5%) can easily lead to overcharge / over-discharge risks. The lack of a multi-device coordination mechanism makes it unable to respond to emergency grid dispatching needs (frequency regulation response time > 500ms).

[0004] 2. Shortcomings of existing improvement technologies Some smart charging solutions are based on a single battery model and do not fully consider the impact of environmental factors (temperature, humidity) on charging performance; The imperfect collaborative architecture between edge computing and the cloud leads to high latency in policy decision-making (≥200ms) and low data security. The hardware devices have low integration and cannot meet the requirements of high-precision and high-reliability industrial applications.

[0005] To address the above problems, this invention proposes an intelligent charging control method, device, and system. Summary of the Invention

[0006] The main objective of this invention is to provide an intelligent charging control method, device, and system that can effectively solve the problems in the background art.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: An intelligent charging control method includes the following specific steps: Step (1) Multi-dimensional data acquisition and fusion: synchronously acquire battery status parameters (SOC±0.5%, SOH±1%, temperature T±0.3℃), grid parameters (electricity price fluctuation ≤5 minutes, load rate ±2%), and environmental parameters (temperature ±0.5℃, humidity ±3% RH); Step (2) Adaptive charging window dynamic generation: Based on the prediction of battery aging trend using Long Short-Term Memory Network (LSTM), the optimal charging parameter window (current I∈[Imin,Imax], voltage U∈[Umin,Umax]) is calculated in real time using Particle Swarm Optimization (PSO) algorithm, with a window update frequency ≥1Hz; Step (3) Multi-objective optimization strategy decision: Construct a three-dimensional objective function that includes charging cost, time efficiency and battery life, solve the Pareto optimal solution set by non-dominated sorting genetic algorithm II (NSGA-II), and output the dynamic charging strategy; Step (4) Distributed collaborative control execution: The collaborative scheduling of the charging pile cluster is realized through edge computing nodes. When the grid load rate is >85%, the graded load reduction mechanism is triggered (Level 1 load reduction of 20%, Level 2 load reduction of 40%), with a response time ≤100ms.

[0008] Preferably, in step (1): the battery status parameters are fused with voltage, current and temperature data through a Kalman filter algorithm, with a sampling frequency ≥10Hz; the power grid parameters are traced through blockchain technology, with a data transmission delay ≤50ms.

[0009] Preferably, in step (2): the LSTM model input layer includes historical charging data (≥1000 cycles), temperature curve (resolution ≤0.1℃), and charge / discharge rate (accuracy ±0.05C); the PSO algorithm has a particle dimension of 5 dimensions (charging current, cutoff voltage, temperature threshold, equalization frequency, pulse duty cycle), and a convergence accuracy ≤0.01.

[0010] Preferably, in step (3), the objective function expression is: Where C is the charging cost, t is the charging time, and N is the battery cycle life. , , The weights are dynamic (summing to 1) and are adjusted in real time through reinforcement learning; the NSGA-II algorithm has a population size of 100, an iteration count of ≥50, and a solution set distribution index GD≤0.05.

[0011] Preferably, in step (4): the edge computing node uses a federated learning mechanism to update the control model, with a local training data volume of ≥10,000 records; the hierarchical load reduction mechanism achieves charging pile cluster synchronization through a consensus algorithm, with a synchronization error of ≤10ms.

[0012] An intelligent charging control device includes: Data acquisition module: integrates a high-precision sensor group (current sensor accuracy ±0.1% FS, temperature sensor response time ≤50ms), a communication interface (supports CAN / LAN / 5G) and a data preprocessing unit (noise suppression ratio ≥40dB). Intelligent decision-making module: equipped with an ARM Cortex-A76 processor (computing power ≥ 20,000 DMIPS), built-in LSTM prediction engine (inference latency ≤ 15ms), PSO optimization module (solution time ≤ 80ms) and policy cache unit (capacity ≥ 1GB). Collaborative control module: includes solid-state relay group (switching time ≤1ms), battery balancing circuit (supports active / passive balancing, voltage balancing accuracy ±5mV) and abnormal handling unit (over-temperature response time ≤20ms).

[0013] Preferably, the data acquisition module further includes a distributed fiber optic temperature sensor (DTS) to achieve single-point temperature monitoring of the battery module (spatial resolution ≤10cm, accuracy ±0.3℃). The power grid data acquisition unit supports IEEE 1588 precise time synchronization (time synchronization error ≤1μs).

[0014] Preferably, the intelligent decision-making module further includes: a multi-objective optimization chip (FPGA architecture) that supports parallel computing of the NSGA-II algorithm (population size ≥ 200, number of iterations ≥ 100). The reinforcement learning unit dynamically adjusts policy weights based on the DDPG algorithm (reward function update frequency ≥ 2Hz).

[0015] An intelligent charging control system, the charging control device as described in any one of claims 6-8; Charging pile cluster (supports charging rates from 0.5C to 3C, communication latency ≤20ms); The battery management system (BMS) features single-cell voltage acquisition (accuracy ±5mV) and thermal runaway early warning functions. The power grid interaction module supports peak-valley electricity price synchronization (update frequency ≤ 5 minutes) and real-time load rate monitoring (accuracy ±1%).

[0016] Preferably, the system is implemented through an edge-cloud collaborative architecture: Edge nodes are responsible for real-time control (latency ≤ 50ms), while the cloud performs policy optimization and data storage (capacity ≥ 10TB). It supports blockchain notarization, with charging data upload latency of ≤2 seconds, meeting the needs of carbon footprint tracking.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. High-precision battery status sensing and real-time monitoring Multi-source data fusion technology: By fusing multi-dimensional data such as voltage, current, and temperature through the Kalman filter algorithm, and combining distributed fiber optic temperature sensors and high-precision current sensors, high-precision monitoring of battery state of charge (SoC), state of health (SoH), and individual cell temperature difference is achieved, which significantly improves the accuracy and real-time performance of battery state estimation and provides a data foundation for precise charging strategies.

[0018] Rapid response to abnormal conditions: Based on real-time temperature field monitoring and intelligent threshold judgment, when the temperature difference of a battery cell exceeds the limit or there is a risk of overheating, the system can quickly trigger active balancing or protection mechanisms to effectively avoid safety hazards such as thermal runaway and improve the safety of the charging process.

[0019] 2. Intelligent Strategy Generation and Multi-Objective Optimization Adaptive dynamic charging window: Utilizes a long short-term memory network (LSTM) to predict battery aging trends, and combines it with a particle swarm optimization (PSO) algorithm to generate the optimal charging parameter window in real time, enabling dynamic adjustment of parameters such as charging current and voltage, while taking into account the speed and safety requirements of different charging stages (such as efficient energy replenishment in the fast charging stage and battery protection in the slow charging stage).

[0020] Multi-objective collaborative optimization: A three-dimensional objective function including charging cost, time efficiency, and battery life is constructed through the non-dominated sorting genetic algorithm (NSGA-II). It supports adaptive switching of strategies in multiple scenarios such as economic mode, fast mode, and lifespan priority mode, which solves the problem that traditional single-objective control cannot take into account multiple needs and improves the overall performance of the charging solution.

[0021] 3. Distributed collaborative control and grid-friendliness improvement Dynamic load balancing of the power grid: Through the coordinated scheduling of edge computing nodes and charging pile clusters, when the power grid load rate exceeds the limit or the frequency fluctuates, a graded load reduction mechanism or power regulation strategy is triggered to realize real-time interaction between charging equipment and the power grid, effectively alleviate the peak and valley pressure of the power grid, and improve energy utilization efficiency and power grid stability.

[0022] Highly efficient collaboration among device clusters: Federated learning and consensus algorithms are used to synchronize parameters and update policies of charging pile clusters, reducing cloud dependence while ensuring high-precision collaboration among devices. The response time is significantly shortened compared to traditional centralized control, meeting the real-time scheduling requirements of large-scale charging networks. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating the intelligent charging control process of an intelligent charging control method, device, and system according to the present invention. Figure 2This is a schematic diagram of the charging control device of the intelligent charging control method, apparatus and system of the present invention. Figure 3 This is a charging control system architecture diagram of an intelligent charging control method, device and system according to the present invention. Detailed Implementation

[0024] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0025] An intelligent charging control method includes the following specific steps: Step (1) Multi-dimensional data acquisition and fusion: Simultaneously acquire battery status parameters (SOC±0.5%, SOH±1%, temperature T±0.3℃), power grid parameters (electricity price fluctuation ≤5 minutes, load rate ±2%), and environmental parameters (temperature ±0.5℃, humidity ±3% RH). The battery status parameters are fused with voltage, current, and temperature data through a Kalman filter algorithm, with a sampling frequency ≥10Hz. The power grid parameters are traced through blockchain technology, with a data transmission delay ≤50ms. Step (2) Dynamic generation of adaptive charging window: Based on the prediction of battery aging trend using Long Short-Term Memory (LSTM) network, the optimal charging parameter window (current I∈[Imin,Imax], voltage U∈[Umin,Umax]) is calculated in real time using the Particle Swarm Optimization (PSO) algorithm. The window update frequency is ≥1Hz. The input layer of the LSTM model contains historical charging data (≥1000 cycles), temperature curve (resolution ≤0.1℃), and charge / discharge rate (accuracy ±0.05C). The particle dimension of the PSO algorithm is 5-dimensional (charging current, cutoff voltage, temperature threshold, equalization frequency, pulse duty cycle), and the convergence accuracy is ≤0.01. Step (3) Multi-objective optimization strategy decision: Construct a three-dimensional objective function that includes charging cost, time efficiency, and battery life. Solve for the Pareto optimal solution set using the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and output the dynamic charging strategy. The objective function expression is: Where C is the charging cost, t is the charging time, and N is the battery cycle life. , , The dynamic weight coefficients (summing to 1) are adjusted in real time through reinforcement learning; the NSGA-II algorithm has a population size of 100, an iteration count ≥ 50, and a solution set distribution index GD ≤ 0.05. Step (4) Distributed collaborative control execution: The collaborative scheduling of the charging pile cluster is realized through edge computing nodes. When the grid load rate is >85%, the graded load reduction mechanism is triggered (Level 1 load reduction of 20%, Level 2 load reduction of 40%), with a response time ≤100ms. The edge computing nodes use the federated learning mechanism to update the control model, with a local training data volume ≥10,000 records. The graded load reduction mechanism realizes the synchronization of the charging pile cluster through the consensus algorithm, with a synchronization error ≤10ms.

[0026] An intelligent charging control device includes: Data acquisition module: integrates a high-precision sensor group (current sensor accuracy ±0.1% FS, temperature sensor response time ≤50ms), a communication interface (supports CAN / LAN / 5G) and a data preprocessing unit (noise suppression ratio ≥40dB). Intelligent decision-making module: equipped with an ARM Cortex-A76 processor (computing power ≥ 20,000 DMIPS), built-in LSTM prediction engine (inference latency ≤ 15ms), PSO optimization module (solution time ≤ 80ms) and policy cache unit (capacity ≥ 1GB). Collaborative control module: includes solid-state relay group (switching time ≤1ms), battery balancing circuit (supports active / passive balancing, voltage balancing accuracy ±5mV) and abnormal handling unit (over-temperature response time ≤20ms). The data acquisition module also includes: Distributed fiber optic temperature sensor (DTS) enables single-point temperature monitoring of battery modules (spatial resolution ≤10cm, accuracy ±0.3℃). The power grid data acquisition unit supports IEEE 1588 precise time synchronization (time synchronization error ≤1μs). The intelligent decision-making module also includes: Multi-objective optimization chip (FPGA architecture) supports parallel computing of NSGA-II algorithm (population size ≥200, number of iterations ≥100). The reinforcement learning unit dynamically adjusts policy weights based on the DDPG algorithm (reward function update frequency ≥ 2Hz).

[0027] An intelligent charging control system, comprising the aforementioned charging control device; Charging pile cluster (supports charging rates from 0.5C to 3C, communication latency ≤20ms); The battery management system (BMS) features single-cell voltage acquisition (accuracy ±5mV) and thermal runaway early warning functions. The power grid interaction module supports peak-valley electricity price synchronization (update frequency ≤ 5 minutes) and real-time load rate monitoring (accuracy ±1%). The system is implemented through an edge-cloud collaborative architecture: Edge nodes are responsible for real-time control (latency ≤ 50ms), while the cloud performs policy optimization and data storage (capacity ≥ 10TB). It supports blockchain notarization, with charging data upload latency of ≤2 seconds, meeting the needs of carbon footprint tracking.

[0028] Example: Scenario: Electric vehicle fast charging station application 1. System Deployment Hardware configuration: Control unit: ARM Cortex-A76 processor + FPGA multi-objective optimization chip; Sensors: Distributed fiber optic temperature sensor (16 channels, covering the battery module), Hall current sensor (accuracy ±0.1% FS); Charging stations: 30 120kW DC fast charging stations (supporting charging rates from 0.5C to 3C).

[0029] Software configuration: Edge nodes: run LSTM prediction models (inference latency ≤ 15ms) and PSO optimization algorithm; Cloud: Deploy the NSGA-II global optimization engine to manage 300 edge nodes.

[0030] 2. Charging process Data acquisition phase (t=0): Battery status: SOC=25%, SOH=88%, module maximum temperature 32℃ (single point temperature difference ≤1℃); Power grid parameters: Peak electricity price 1.2 yuan / kWh, load factor 87%; Environmental parameters: ambient temperature 28℃, humidity 65% ​​RH.

[0031] Strategy generation phase (t=1-2s): LSTM prediction: Capacity decay rate of 2.1% in the next 2 hours (actual value 2.3%, error 8.7%). PSO optimization: Optimal charging parameters I=120A (1.5C), U=4.18V, temperature threshold 38℃; NSGA-II Decision: Choose a compromise strategy (cost weight 0.4, time weight 0.3, lifespan weight 0.3), with an estimated charging time of 40 minutes, cost of 57.6 yuan, and battery degradation of 0.04%.

[0032] Collaborative control phase: t=10min: The grid load rate rises to 89%, triggering Level 1 load descent (current drops to 96A), and the system automatically pushes a compensation plan (extending the off-peak charging time by 3 minutes). At t=25min: the temperature difference between individual battery cells is 4.5℃ (close to the 5℃ threshold), the equalization circuit is activated, and the current difference between individual cells is adjusted to ≤0.4A; t=42min: Charging complete (SOC=80%), actual cost 62.4 yuan (error + 8.3%), maximum battery temperature 36.5℃ (<38℃ threshold).

[0033] 3. Comparative Experiment index Traditional fast charging This invention difference Charging time 48 minutes 42 minutes -12.5% Charging costs 72 yuan 62.4 yuan -13.3% Battery temperature rise 8℃ 4.5℃ -43.8% Individual temperature difference 3.2mV 0.8mV -75% The above embodiments and data demonstrate that the present invention provides an intelligent charging control method, device, and system. Through multi-dimensional data fusion, intelligent algorithm optimization, and hardware device innovation, it achieves high efficiency, safety, and grid-friendly charging process, demonstrating significant technological advancement and application value.

[0034] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of this invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent charging control method, characterized in that: The specific steps include the following: Step (1) Multi-dimensional data acquisition and fusion: synchronously acquire battery status parameters (SOC±0.5%, SOH±1%, temperature T±0.3℃), power grid parameters (electricity price fluctuation ≤5 minutes, load rate ±2%), and environmental parameters (temperature ±0.5℃, humidity ±3% RH); Step (2) Adaptive charging window dynamic generation: Based on the prediction of battery aging trend using Long Short-Term Memory Network (LSTM), the optimal charging parameter window (current I∈[Imin,Imax], voltage U∈[Umin,Umax]) is calculated in real time using Particle Swarm Optimization (PSO) algorithm, with a window update frequency ≥1Hz; Step (3) Multi-objective optimization strategy decision: Construct a three-dimensional objective function that includes charging cost, time efficiency and battery life, solve the Pareto optimal solution set by non-dominated sorting genetic algorithm II (NSGA-II), and output the dynamic charging strategy; Step (4) Distributed collaborative control execution: The collaborative scheduling of the charging pile cluster is realized through edge computing nodes. When the grid load rate is >85%, the graded load reduction mechanism is triggered (Level 1 load reduction of 20%, Level 2 load reduction of 40%), with a response time ≤100ms.

2. The intelligent charging control method according to claim 1, characterized in that: In step (1): the battery status parameters are fused with voltage, current and temperature data through Kalman filtering algorithm, with a sampling frequency ≥10Hz; the power grid parameters are traced through blockchain technology, with a data transmission delay ≤50ms.

3. The intelligent charging control method according to claim 1, characterized in that: In step (2): the input layer of the LSTM model contains historical charging data (≥1000 cycles), temperature curves (resolution ≤0.1℃), and charge / discharge rate (accuracy ±0.05C); the particle dimension of the PSO algorithm is 5-dimensional (charging current, cutoff voltage, temperature threshold, equalization frequency, pulse duty cycle), and the convergence accuracy is ≤0.

01.

4. The intelligent charging control method according to claim 1, characterized in that: In step (3), the objective function expression is: Where C is the charging cost, t is the charging time, and N is the battery cycle life. , , The weights are dynamic (summing to 1) and are adjusted in real time through reinforcement learning; the NSGA-II algorithm has a population size of 100, an iteration count of ≥50, and a solution set distribution index GD≤0.

05.

5. The intelligent charging control method according to claim 1, characterized in that: In step (4): the edge computing node uses a federated learning mechanism to update the control model, with a local training data volume of ≥10,000 records; the hierarchical load reduction mechanism achieves charging pile cluster synchronization through a consensus algorithm, with a synchronization error of ≤10ms.

6. An intelligent charging control device, characterized in that: include: Data acquisition module: integrates a high-precision sensor group (current sensor accuracy ±0.1% FS, temperature sensor response time ≤50ms), a communication interface (supports CAN / LAN / 5G) and a data preprocessing unit (noise suppression ratio ≥40dB). Intelligent decision-making module: equipped with an ARM Cortex-A76 processor (computing power ≥20,000 DMIPS), and built-in LSTM prediction engine (inference latency ≤15ms), PSO optimization module (solution time ≤80ms) and policy cache unit (capacity ≥1GB). Collaborative control module: includes solid-state relay group (switching time ≤1ms), battery balancing circuit (supports active / passive balancing, voltage balancing accuracy ±5mV) and abnormal handling unit (over-temperature response time ≤20ms).

7. The intelligent charging control device according to claim 6, characterized in that: The data acquisition module also includes: Distributed fiber optic temperature sensor (DTS) enables single-point temperature monitoring of battery modules (spatial resolution ≤10cm, accuracy ±0.3℃). The power grid data acquisition unit supports IEEE 1588 precise time synchronization (time synchronization error ≤1μs).

8. The intelligent charging control device according to claim 6, characterized in that: The intelligent decision-making module also includes: Multi-objective optimization chip (FPGA architecture) supports parallel computing of NSGA-II algorithm (population size ≥200, number of iterations ≥100). The reinforcement learning unit dynamically adjusts policy weights based on the DDPG algorithm (reward function update frequency ≥ 2Hz).

9. An intelligent charging control system, characterized in that: The charging control device as described in any one of claims 6-8; Charging pile cluster (supports charging rates from 0.5C to 3C, communication latency ≤20ms); The battery management system (BMS) features single-cell voltage acquisition (accuracy ±5mV) and thermal runaway early warning functions. The power grid interaction module supports peak-valley electricity price synchronization (update frequency ≤ 5 minutes) and real-time load rate monitoring (accuracy ±1%).

10. The intelligent charging control system according to claim 9, characterized in that: The system is implemented through an edge-cloud collaborative architecture: Edge nodes are responsible for real-time control (latency ≤ 50ms), while the cloud performs policy optimization and data storage (capacity ≥ 10TB). It supports blockchain notarization, with charging data upload latency of ≤2 seconds, meeting the needs of carbon footprint tracking.

Images