Intelligent charging and discharging regulation and control system and method
By predicting user travel patterns using deep learning models and combining them with multi-gun collaborative control strategies, the timing of V2G charging and discharging is optimized. This solves the problems of power grid and vehicle demand balance and low user participation in the existing system, and achieves two-way optimization of power grid peak shaving and valley filling and user travel guarantee.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-14
- Publication Date
- 2026-04-03
AI Technical Summary
Existing V2G charging and discharging systems lack bidirectional charging and discharging coordination control capabilities in multi-gun scenarios, have low power module switching efficiency, cannot take into account the dynamic balance between grid load and vehicle demand, and rely on a single electricity price signal for charging and discharging timing, lacking quantitative assessment of travel security, resulting in low user participation.
It uses a deep learning model to predict user travel patterns, determines the charging and discharging operation window, and optimizes the timing of charging and discharging through a multi-gun collaborative control strategy. Combined with grid load and vehicle SOC, it dynamically adjusts the switching of power modules to provide accurate charging and discharging reminders and revenue predictions.
It has achieved two-way optimization of power grid peak shaving and valley filling and user travel guarantee, improved the intelligence and efficiency of V2G system, and increased users' willingness to participate in V2G.
Smart Images
Figure CN121791237A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to an intelligent charging and discharging control system and method, belonging to the field of V2G charging control technology. Background Technology
[0002] With the increasing popularity of electric vehicles, V2G technology has become a key means to achieve peak shaving and valley filling in the power grid and the consumption of renewable energy. Existing V2G charging and discharging systems have the following shortcomings: First, multi-gun charging piles mainly focus on charging power allocation, lacking bidirectional charging and discharging coordination control capabilities in scenarios with ≥3 guns, and the power module switching efficiency is low, making it difficult to balance the dynamic equilibrium between grid load and vehicle demand; second, charging and discharging timing reminders rely heavily on a single electricity price signal, completely neglecting user travel plans, easily leading to insufficient battery power or revenue loss; third, the lack of a quantitative travel guarantee assessment mechanism limits user willingness to participate in V2G. In summary, the current V2G charging and discharging control process cannot simultaneously achieve grid peak shaving and valley filling and user travel guarantee, resulting in relatively low user participation. Summary of the Invention
[0003] The purpose of this invention is to provide an intelligent charging and discharging control system and method to solve the problem that the current V2G charging control process cannot simultaneously address the issues of grid peak shaving and valley filling and ensuring user travel safety.
[0004] To solve the above-mentioned technical problems, the present invention provides an intelligent charge and discharge control method, the method comprising: 1) Obtain the user's expected departure time and expected travel mileage based on the user's travel pattern prediction results or travel plan, and determine the user's charging and discharging operation window based on the expected departure time; 2) Determine whether charging is needed based on the user's remaining battery power. If charging is needed, start charging first. Otherwise, determine the travel guarantee level based on the user's remaining battery power range and expected travel range. If the travel guarantee level meets the user's needs, send a message to the user at the beginning of the user's charging / discharging operation window. The message includes a discharging operation suggestion and the operation duration. If the travel guarantee level does not meet the user's needs, charge for a period of time first, and recalculate the travel guarantee level after charging to determine if it meets the user's needs. If it does, immediately send a message to the user containing a discharging operation suggestion. Otherwise, continue charging and repeat the process.
[0005] Furthermore, the user travel pattern prediction results are obtained based on deep learning model training, and the training data is the user's historical travel data, including the user's departure time and travel mileage each time.
[0006] Furthermore, the deep learning model described uses the LSTM model.
[0007] Furthermore, when the message being sent is a discharge operation, the method also calculates the net benefit based on the peak-valley voltage difference of the power grid and the discharge subsidy standard, and sends the net benefit as part of the message to the user.
[0008] Furthermore, the charging station for charging and discharging is equipped with multiple charging guns, and each charging gun adopts the following collaborative control strategy: Collect the real-time load of the power grid and the SOC and rated power of the vehicles connected to each charging gun; The theoretical optimal power of each charging gun is calculated based on the SOC and rated power of the connected vehicle. The number of power modules that need to be switched on for each charging gun is determined based on the grid load, the theoretical optimal power of each charging gun, and the load rate of each power module. The corresponding switching is completed according to the determined number of power modules to meet the grid load and user needs.
[0009] Furthermore, when switching power modules, the output current of the power module to be switched is first controlled to drop to 0A, and then the corresponding switching circuit is triggered to switch, so as to achieve safe switching of power modules.
[0010] Furthermore, when coordinating and controlling each charging gun, the collected real-time grid load and vehicle SOC data are updated at set intervals. The power allocation scheme is dynamically corrected based on the updated data to ensure that each charging gun is always in the optimal charging and discharging state.
[0011] Furthermore, the travel security level is the ratio of the user's remaining battery power range to the expected travel range. A travel security level that meets the user's needs means that the travel security level is greater than a set threshold.
[0012] Furthermore, when all charging guns are controlled in a coordinated manner, the power module prioritizes allocating power to the charging guns connected to the vehicle with the lower SOC.
[0013] The present invention also provides an intelligent charge-discharge regulation system, including a processor, the processor being used to execute a relevant computer program to implement an intelligent charge-discharge regulation method, the method comprising: 1) Obtain the user's expected departure time and expected travel mileage based on the user's travel pattern prediction results or travel plan, and determine the user's charging and discharging operation window based on the expected departure time; 2) Determine whether charging is needed based on the user's remaining battery power. If charging is needed, start charging first. Otherwise, determine the travel guarantee level based on the user's remaining battery power range and expected travel range. If the travel guarantee level meets the user's needs, send a message to the user at the beginning of the user's charging / discharging operation window. The message includes a discharging operation suggestion and the operation duration. If the travel guarantee level does not meet the user's needs, charge for a period of time first, and recalculate the travel guarantee level after charging to determine if it meets the user's needs. If it does, immediately send a message to the user containing a discharging operation suggestion. Otherwise, continue charging and repeat the process.
[0014] Furthermore, the user travel pattern prediction results are obtained based on deep learning model training, and the training data is the user's historical travel data, including the user's departure time and travel mileage each time.
[0015] Furthermore, the deep learning model described herein employs the LSTM model.
[0016] Furthermore, when the message being sent is a discharge operation, the method also calculates the net benefit based on the peak-valley voltage difference of the power grid and the discharge subsidy standard, and sends the net benefit as part of the message to the user.
[0017] Furthermore, the charging station for charging and discharging is equipped with multiple charging guns, and each charging gun adopts the following collaborative control strategy: Collect the real-time load of the power grid and the SOC and rated power of the vehicles connected to each charging gun; The theoretical optimal power of each charging gun is calculated based on the SOC and rated power of the connected vehicle. The number of power modules that need to be switched on for each charging gun is determined based on the grid load, the theoretical optimal power of each charging gun, and the load rate of each power module. The corresponding switching is completed according to the determined number of power modules to meet the grid load and user needs.
[0018] Furthermore, when switching power modules, the output current of the power module to be switched is first controlled to drop to 0A, and then the corresponding switching circuit is triggered to switch, so as to achieve safe switching of power modules.
[0019] Furthermore, when coordinating and controlling each charging gun, the collected real-time grid load and vehicle SOC data are updated at set intervals. The power allocation scheme is dynamically corrected based on the updated data to ensure that each charging gun is always in the optimal charging and discharging state.
[0020] Furthermore, the travel security level is the ratio of the user's remaining battery power range to the expected travel range. A travel security level that meets the user's needs means that the travel security level is greater than a set threshold.
[0021] Furthermore, when all charging guns are controlled in a coordinated manner, the power module prioritizes allocating power to the charging guns connected to the vehicle with the lower SOC.
[0022] The beneficial effects of this invention are as follows: This invention determines the travel security level based on the user's remaining battery power and expected travel mileage. When the travel security level meets the user's needs, it sends a message to the user at the start of the user's charging / discharging operation window, including charging / discharging operation suggestions and operation duration. Therefore, this invention obtains the user's expected departure time and expected travel mileage, and determines the user's charging / discharging operation window based on the expected departure time and charging / discharging power. Under the premise of meeting the travel security level, it pushes charging / discharging opportunities within a precise time window, thereby achieving two-way optimization of power grid peak shaving and valley filling and user travel security. This improves the intelligence and efficiency of the V2G system and increases users' willingness to participate in V2G. Attached Figure Description
[0023] Figure 1 This is a schematic diagram of the architecture of the intelligent charge and discharge control system in an embodiment of the present invention; Figure 2 This is a flowchart of the multi-gun load coordination control in an embodiment of the present invention; Figure 3 This is a flowchart illustrating the precise charging and discharging reminder process in an embodiment of the present invention; Figure 4 This is a schematic diagram of the LSTM model used for travel pattern prediction in an embodiment of the present invention. Detailed Implementation
[0024] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.
[0025] This invention obtains the user's expected departure time and expected travel mileage, and determines the user's charging and discharging operation window based on the expected departure time and charging and discharging power. Under the premise of ensuring travel security, it pushes charging and discharging timing and revenue prediction within a precise time window, so as to achieve two-way optimization of power grid peak shaving and valley filling and user travel security, thereby improving the intelligence and efficiency of the V2G system.
[0026] Implementation of the Intelligent Charge and Discharge Regulation Method of the Invention The intelligent charging and discharging control method of this invention includes two aspects of control: charging and discharging reminders and multi-gun load coordination control. The charging and discharging reminders are based on the user's estimated departure time and estimated travel mileage. First, the user's estimated departure time and estimated travel mileage are obtained. Based on the estimated departure time, estimated travel mileage, and charging / discharging power, the user's charging and discharging operation window is determined. The travel guarantee level is determined based on the user's remaining battery power and estimated travel mileage. When the travel guarantee level meets the user's needs, a message is sent to the user at the start of the user's charging and discharging operation window. The specific implementation process is as follows: Figure 3 As shown. Multi-barrel load coordination control calculates the theoretical optimal power of each charging barrel, determines the number of power modules that need to be switched on for each charging barrel based on the theoretical optimal power of each charging barrel, and completes the corresponding switching according to the determined number of power modules to meet the charging needs of each charging barrel. The specific implementation process is as follows. Figure 2 As shown below. The two processes described above will be explained in detail below.
[0027] Before providing a detailed description of the intelligent charge and discharge control method of the present invention, the charging method employed in the present invention will now be explained. The present invention employs V2G charging and discharging technology, which achieves bidirectional power exchange between the power grid and the charging terminal through a power conversion unit, such as... Figure 1 As shown, the power conversion unit consists of a DC-DC module and a V2G bidirectional module. The V2G bidirectional module enables bidirectional energy transfer between the AC grid and the DC side, while the DC-DC module matches the vehicle battery voltage level and supports a wide range of voltage regulation.
[0028] I. Charge / Discharge Reminder Control 1. Obtain the user's expected departure time and expected travel distance.
[0029] The user's estimated departure time and estimated travel mileage can be obtained from the user's travel pattern prediction results or travel plan. When there is a user travel plan, it can be directly obtained from the travel plan; when there is no user travel plan, prediction can only be made based on the user's travel patterns. Specifically, this invention uses a deep learning model as the prediction model, and trains the prediction model using the user's historical travel pattern data, thus enabling the prediction of user travel patterns. This embodiment uses an LSTM model as the prediction model. During model training, historical user travel data is used, for example, one day at a time, collecting the user's travel data from the past year as training data. The travel data includes departure time and travel mileage. The user's departure time and travel mileage from the past year are input into the LSTM model for training, and the LSTM model can find the user's travel patterns. Thus, in practical applications, inputting the user's departure time and travel mileage from the past 30 days into the trained LSTM model can obtain the departure time and travel mileage for the next day.
[0030] This implementation uses LSTM as the architecture for the prediction model, as follows: Figure 4 As shown, the system includes an input feature layer, a forget gate, an input gate, an output gate, and a prediction output layer. The input feature layer takes three features as input: Feature 1: daily departure time for the past 30 days; Feature 2: daily commuting distance for the past 30 days; and Feature 3: daily parking time for the past 30 days (e.g., 8 hours). Feature 3 is a prediction reference and may be omitted. Alternatively, other deep learning models, such as deep convolutional neural network models, can be used to predict user travel data.
[0031] 2. Determine the user's charging and discharging operation window, and determine the travel guarantee level based on the user's remaining battery power range and expected travel range.
[0032] The charging and discharging operation window is calculated by working backward from the estimated departure time (generally 4-8 hours before departure). The charging and discharging operation window is also calculated by combining the charging and discharging power and the vehicle's current SOC (remaining battery power) to ensure that the charging and discharging operation is completed within the charging and discharging operation window.
[0033] Specifically, the system calculates whether charging / discharging is feasible based on the estimated departure time and the vehicle's current State of Charge (SOC). If charging is required, it is done promptly. If the SOC is sufficient to meet travel needs and there is still some capacity remaining, a strategy is implemented to determine whether to discharge. The user's remaining battery power can be obtained from the Battery Management System (BMS) data. Based on this remaining battery power, the remaining driving range can be calculated. The remaining battery power represents the vehicle's SOC before charging / discharging. The travel security level is calculated by comparing the remaining driving range with the estimated travel range; the travel security level is the ratio of the remaining driving range to the estimated travel range. This implementation uses the travel security level to measure whether travel needs are met.
[0034] 3. Control reminder push notifications based on travel security levels.
[0035] When the user's travel security level meets their needs, a message is sent to the user at the start of the charging / discharging operation window. Meeting user needs means the travel security level is greater than a set threshold, which is a user-defined value, such as 90%. When the calculated travel security level is greater than 90%, a reminder is pushed to the user's terminal at the start of the charging / discharging operation window. The reminder includes operation suggestions (charging / discharging), estimated operation time, and travel security level. To allow users to intuitively understand the benefits of the current operation suggestions, the calculated net benefit is also sent to the user's terminal as part of the reminder. The net benefit is calculated based on the peak-valley voltage difference of the power grid and the discharge subsidy standard: Net benefit = Discharge benefit - Charging cost + Subsidy. The net benefit is calculated only if a discharge operation is mentioned; if no discharge operation is mentioned in the reminder, no net benefit is calculated. Sending a message at the start of the user's charging / discharging operation window informs the customer that the operation has begun. If the customer has other needs, they can stop the operation, thus increasing the flexibility of control.
[0036] When the calculated travel security level does not meet the requirements (e.g., not greater than 90%), a charging operation is performed for a period of time. After the charging operation lasts for a period of time (e.g., 1 hour), the travel security level is recalculated. The user's needs are then determined based on the new travel security level. If the travel security level meets the requirements, the user can proceed with the discharge operation and receive a corresponding reminder push. Otherwise, the charging operation and judgment continue to be performed. This is a cyclical process that continues until the requirements are met or the charging and discharging operation window period is exceeded.
[0037] Through the above-mentioned controls, the present invention can reduce the risk of insufficient battery power during travel while meeting economic requirements, thereby increasing users' willingness to participate in V2G.
[0038] II. Multi-gun load coordination control When a charging station is equipped with multiple charging guns (3 or more), it is necessary to consider the grid load and the load demand of each charging gun to implement coordinated control of each charging gun in order to balance the grid demand and the interests of the user. The implementation process of this coordinated control is as follows: Figure 2 As shown below, a detailed explanation will follow.
[0039] 1. Obtain the real-time load of the power grid and the load demand of each charging gun.
[0040] This invention utilizes the energy centralized management unit on the charging pile side to collect the load of the power grid and the connection status of each charging gun in real time. When a vehicle is connected to the charging gun, the vehicle's SOC and rated charging and discharging power are obtained, and a decision is made based on the vehicle's SOC and rated charging and discharging power.
[0041] 2. Control the switching of power modules according to the real-time load of the power grid and the load demand of each charging gun.
[0042] like Figure 1 As shown, the charging pile contains multiple power modules. A certain number of power modules need to be allocated to each charging gun based on its load demand. If the grid load is insufficient, full-power charging can be performed; if the grid load is excessive, charging at reduced power or a discharge strategy should be implemented. For example, if a single power module has a power of 100W, and a charging gun has a load demand of 240W, then without considering the grid load, three power modules need to be allocated to that charging gun.
[0043] Specifically, the theoretical optimal power of each charging gun is calculated based on its load demand (determined by the SOC of the connected vehicle) and rated power. This theoretical optimal power refers to the maximum power required to meet the charging gun's demand. The number of power modules needed for each charging gun is determined based on the theoretical optimal power and the load rate of the power modules. The corresponding switching is then performed according to the determined number of power modules to meet the charging needs of each charging gun. The power module achieves the highest efficiency when its load rate is between 50% and 80%. Therefore, when determining the number of power modules, the load rate should be considered, aiming to keep it between 50% and 80%. For example, if a power module has a power of 100W and a charging gun requires 200W, then three power modules are needed for that charging gun, ensuring that the load rate of these three power modules is between 50% and 80%. Theoretically, charging power is allocated on a first-come, first-served basis. However, when a vehicle's State of Charge (SOC) exceeds 80%, the Battery Management System (BMS) will proactively reduce the charging current request, thus decreasing the charging power demand. The remaining power will then be allocated to vehicles with lower SOCs. Furthermore, in multi-gun concurrent charging scenarios, the charging pile's power allocation strategy prioritizes the needs of vehicles with lower SOCs, relatively mitigating the impact of the first-come, first-served principle. While the first-come, first-served principle is the fundamental rule for charging pile power allocation, the power reduction at high SOCs is an inevitable result of the combined effect of battery protection mechanisms and intelligent scheduling strategies. For example, when four vehicles are connected simultaneously, the grid load is 300kW (below the 336kW threshold). The system allocates a total power of 480kW based on each vehicle's rated power (total charging pile power 480kW, including eight 60kW power modules). When the grid load rises to 400kW, the system adjusts to prioritize charging two low-SOC vehicles (total 200kW) and discharging two high-SOC vehicles (total 100kW), keeping grid load fluctuations within ±5%.
[0044] Implementation of the Intelligent Charge and Discharge Control System of the Invention The intelligent charge-discharge regulation system of this invention includes a processor, which executes relevant computer programs to implement an intelligent charge-discharge regulation method. Specifically, the intelligent charge-discharge regulation system of this invention can be divided into an intelligent decision-making unit and an energy centralized management unit as needed, such as... Figure 1As shown, this intelligent decision-making unit is built on an industrial-grade server and incorporates a pre-trained LSTM travel prediction model and a multi-dimensional scoring model. The LSTM model, trained on historical travel data, achieves a prediction accuracy of ≥92%. The energy centralized management unit collects vehicle BMS data from each charging gun via the CAN bus and obtains grid load data via power line carrier communication. It dynamically allocates total power based on a preset algorithm and supports 4G and Bluetooth communication. On one hand, it reads BMS data such as vehicle SOC and battery health; on the other hand, it obtains the user's departure time and estimated mileage for the next day's trip from the navigation app or from the intelligent decision-making unit. The intelligent decision-making unit also communicates with user terminals, including a mobile app and a charging pile touchscreen, supporting the display of security ratings, revenue predictions, and receiving reminder information, and providing a threshold setting interface. Furthermore, as... Figure 1 As shown, the energy centralized management unit is located on the charging pile side, and the charging pile side is also equipped with a power conversion unit, which consists of a DC-DC module and a V2G bidirectional module. One side of the charging pile is connected to the power grid, and the other side is connected to the charging terminal. The charging terminal is equipped with multiple charging guns and a charge and discharge controller, as well as a billing unit.
[0045] The intelligent decision-making unit and the centralized energy management unit can realize charging and discharging reminder control and multi-gun load coordination control. The following is a detailed description of these two control methods.
[0046] I. Charge / Discharge Reminder Control 1. Obtain the user's expected departure time and expected travel distance.
[0047] The user's estimated departure time and estimated travel mileage can be obtained from the user's travel pattern prediction results or travel plan. When there is a user travel plan, it can be directly obtained from the travel plan; when there is no user travel plan, prediction can only be made based on the user's travel patterns. Specifically, this invention uses a deep learning model as the prediction model, and trains the prediction model using the user's historical travel pattern data, thus enabling the prediction of user travel patterns. This embodiment uses an LSTM model as the prediction model. During model training, historical user travel data is used, for example, one day at a time, collecting the user's travel data from the past year as training data. The travel data includes departure time and travel mileage. The user's departure time and travel mileage from the past year are input into the LSTM model for training, and the LSTM model can find the user's travel patterns. Thus, in practical applications, inputting the user's departure time and travel mileage from the past 30 days into the trained LSTM model can obtain the departure time and travel mileage for the next day.
[0048] 2. Determine the user's charging and discharging operation window, and determine the travel guarantee level based on the user's remaining battery power range and expected travel range.
[0049] The charging and discharging operation window is calculated by working backward from the estimated departure time (generally 4-8 hours before departure). The charging and discharging operation window is also calculated by combining the charging and discharging power and the vehicle's current SOC (remaining battery power) to ensure that the charging and discharging operation is completed within the charging and discharging operation window.
[0050] Specifically, the system calculates whether charging or discharging is possible based on the expected departure time and the current vehicle SOC. If charging is required, it charges promptly. If the SOC is sufficient to meet travel needs and there is still some capacity remaining, the system decides whether to discharge based on the strategy.
[0051] The user's remaining battery power can be obtained from the BMS data. Based on the remaining battery power, the remaining driving range can be calculated. Here, the remaining battery power is the vehicle's State of Charge (SOC) before charging and discharging. By using the remaining driving range and the obtained estimated travel range, the travel security level can be calculated. The travel security level is the ratio of the remaining driving range to the estimated travel range. This implementation uses the travel security level to measure whether travel needs are met.
[0052] 3. Control reminder push notifications based on travel security levels.
[0053] When the travel security level meets the user's needs, a message is sent to the user at the start of the user's charging / discharging operation window. Meeting the user's travel security level means that the travel security level is greater than a set threshold, which is a user-defined value, such as 90%. When the calculated travel security level is greater than 90%, a reminder is pushed to the user's terminal at the start of the charging / discharging operation window. The reminder includes operation suggestions (charging / discharging), estimated operation time, and travel security level. To allow users to intuitively understand the benefits of the current discharging operation suggestion, the calculated net benefit is also sent to the user's terminal as part of the reminder. The net benefit is calculated based on the peak-valley voltage difference of the power grid and the discharging subsidy standard: Net Benefit = Discharging Benefit - Charging Cost + Subsidy. The net benefit is calculated only if a discharging operation is performed; if the reminder does not mention a discharging operation, no net benefit is calculated.
[0054] When the calculated travel security level does not meet the requirements (e.g., not greater than 90%), a charging operation is performed for a period of time. After a charging operation period (e.g., 1 hour), the travel security level is recalculated. The user's needs are then determined based on the new travel security level. If the travel security level meets the requirements, the user can proceed with the discharging operation and receive a corresponding reminder push. Otherwise, the charging operation and judgment continue. This is a cyclical process that continues until the requirements are met or the charging / discharging operation window period is exceeded.
[0055] Through the above-mentioned controls, the present invention can reduce the risk of insufficient battery power during travel while meeting economic requirements, thereby increasing users' willingness to participate in V2G.
[0056] II. Multi-gun load coordination control When a charging station is equipped with multiple charging guns (3 or more), it is necessary to consider the grid load and the load demand of each charging gun to implement coordinated control of each charging gun in order to balance the grid demand and the interests of the user. The implementation process of this coordinated control is as follows: Figure 2 As shown below, a detailed explanation will follow.
[0057] 1. Obtain the real-time load of the power grid and the load demand of each charging gun.
[0058] This invention utilizes the energy centralized management unit on the charging pile side to collect the load of the power grid and the connection status of each charging gun in real time. When a vehicle is connected to the charging gun, the vehicle's SOC and rated charging and discharging power are obtained, and a decision is made based on the vehicle's SOC and rated charging and discharging power.
[0059] 2. Control the switching of power modules according to the real-time load of the power grid and the load demand of each charging gun.
[0060] like Figure 1As shown, the charging pile contains multiple power modules. A certain number of power modules need to be allocated to each charging gun based on its load demand. If the grid load is insufficient, full-power charging can be performed; if the grid load is excessive, charging power must be reduced or a discharge strategy must be implemented. For example, if a single power module has a power of 100W, and a charging gun has a load demand of 240W, then, without considering the grid load, three power modules need to be allocated to that charging gun. Specifically, the theoretical optimal power of each charging gun is calculated based on its load demand (determined by the SOC of the connected vehicle) and rated power. The theoretical optimal power refers to the maximum power that meets the charging gun's demand. The number of power modules that need to be switched on for each charging gun is determined based on the theoretical optimal power of each charging gun and the load rate of the power modules. The corresponding switching is then performed according to the determined number of power modules to meet the charging demand of each charging gun. Power modules are most efficient when their load rate is between 50% and 80%. This means that when determining the number of power modules, their load rate should be considered, ideally kept between 50% and 80%. For example, if a power module has a power output of 100W and a charging gun requires 200W, then three power modules should be configured for that charging gun, ensuring their load rate is between 50% and 80%. Theoretically, charging power is allocated on a first-come, first-served basis. However, when a vehicle's State of Charge (SOC) exceeds 80%, the Battery Management System (BMS) will proactively reduce the charging current request, decreasing the charging power demand. The remaining power will then be allocated to vehicles with lower SOCs. Furthermore, in multi-gun concurrent charging scenarios, the charging station's power allocation strategy prioritizes vehicles with lower SOCs, relatively mitigating the impact of the first-come, first-served principle. While the first-come, first-served principle is the fundamental rule for charging station power allocation, the power reduction at high SOCs is an inevitable result of the combined effect of battery protection mechanisms and intelligent scheduling strategies.
[0061] The following is a specific example: deploying this system in a charging station in an urban complex, configuring one charging stack (total power 480kW, including eight 60kW power modules), four charging guns, an energy centralized management unit using an STM32F407 processor, and an intelligent decision-making unit using a Huawei Atlas200AI server.
[0062] 1. Model Training: One year of travel data from 500 users (a total of 182,500 records) was collected, and the training set and test set were divided in a 7:3 ratio to train the LSTM model. The final prediction accuracy reached 94.3%.
[0063] 2. Actual operation: When 4 vehicles are connected at the same time, the grid load is 300kW (below the 336kW threshold), and the system allocates a total power of 480kW according to the rated power of each vehicle; when the grid load rises to 400kW, the system adjusts to prioritize charging 2 low SOC vehicles (total 200kW) and discharging 2 high SOC vehicles (total 100kW), and the grid load fluctuation is controlled within ±5%.
[0064] 3. Push notification: The user departs at 8:00 the next day, with an estimated mileage of 50km. The system predicts the window period to be from 0:00 to 4:00. After charging and discharging, the remaining mileage is calculated to be 62km, and the protection score is 124%. A discharge reminder is pushed, showing a net benefit of 25.6 yuan.
[0065] This invention calculates travel security based on LSTM prediction results and provides charging / discharging reminders accordingly, reducing the risk of insufficient battery power and increasing user willingness to participate in V2G. Simultaneously, through peak shaving and valley filling, the average annual V2G revenue per vehicle increases, while reducing the peak-valley difference in the power grid, alleviating power grid peak-shaving pressure, and increasing average annual carbon reduction. Furthermore, through simultaneous charging / discharging control of multiple charging guns and dynamic power allocation, the total power utilization rate of the charging pile is increased to over 95%, effectively increasing the number of vehicles served and reducing grid load fluctuations compared to traditional dual-gun systems.
Claims
1. An intelligent charging and discharging control method, characterized in that, The method includes: 1) Obtain the user's expected departure time and expected travel mileage based on the user's travel pattern prediction results or travel plan, and determine the user's charging and discharging operation window based on the expected departure time; 2) Determine whether charging is needed based on the user's remaining battery power. If charging is needed, start charging first. Otherwise, determine the travel guarantee level based on the user's remaining battery power range and expected travel range. If the travel guarantee level meets the user's needs, send a message to the user at the beginning of the user's charging / discharging operation window. The message includes a discharging operation suggestion and the operation duration. If the travel guarantee level does not meet the user's needs, charge for a period of time first, and recalculate the travel guarantee level after charging to determine if it meets the user's needs. If it does, immediately send a message to the user containing a discharging operation suggestion. Otherwise, continue charging and repeat the process.
2. The intelligent charge / discharge control method according to claim 1, characterized in that, The user travel pattern prediction results are obtained based on deep learning model training. The training data is the user's historical travel data, including the user's departure time and travel distance for each trip.
3. The intelligent charge / discharge control method according to claim 2, characterized in that, The deep learning model described uses the LSTM model.
4. The intelligent charge / discharge control method according to claim 1, characterized in that, When the message sent is a discharge operation, the method also calculates the net benefit based on the peak-valley voltage difference of the power grid and the discharge subsidy standard, and sends the net benefit to the user as part of the message sent.
5. The intelligent charge / discharge control method according to claim 1, characterized in that, The charging station for charging and discharging is equipped with multiple charging guns, and each charging gun adopts the following cooperative control strategy: Collect the real-time load of the power grid and the SOC and rated power of the vehicles connected to each charging gun; The theoretical optimal power of each charging gun is calculated based on the SOC and rated power of the connected vehicle. The number of power modules that need to be switched on for each charging gun is determined based on the grid load, the theoretical optimal power of each charging gun, and the load rate of each power module. The corresponding switching is completed according to the determined number of power modules to meet the grid load and user needs.
6. The intelligent charge and discharge control method according to claim 5, characterized in that, When switching power modules, the output current of the power module to be switched is first reduced to 0A, and then the corresponding switching circuit is triggered to switch it, so as to achieve safe switching of power modules.
7. The intelligent charge / discharge control method according to claim 5, characterized in that, When coordinating and controlling each charging gun, the collected real-time grid load and vehicle SOC data are updated at set intervals. The power allocation scheme is dynamically adjusted based on the updated data to ensure that each charging gun is always in the optimal charging and discharging state.
8. The intelligent charge and discharge control method according to claim 1, characterized in that, Travel security is the ratio of the user's remaining battery power range to the expected travel range. Travel security meeting user needs means that the travel security is greater than the set threshold.
9. The intelligent charge / discharge control method according to claim 5, characterized in that, When all charging guns are controlled in a coordinated manner, the power module prioritizes allocating power to the charging guns connected to the vehicle with the lower SOC.
10. An intelligent charge / discharge control system, comprising a processor, characterized in that, The processor is used to execute relevant computer programs to implement the intelligent charge and discharge control method according to any one of claims 1-9.