Electric vehicle bidirectional charging power grid load intelligent scheduling method, system and device based on artificial intelligence, and storage medium

Abstract

The invention provides an electric vehicle bidirectional charging power grid load intelligent scheduling method, system and device based on artificial intelligence, and a storage medium, and the method comprises the steps: regularly collecting power grid load data in a preset time period of a preset power utilization region; the power grid load data comprises first load data, electric vehicle charging demand data, user behavior data and weather data; performing data cleaning and feature extraction on the power grid load data, and performing data fusion to generate a fused data set; the first algorithm model processes the fused data set to obtain load prediction data of the power grid; according to the power grid load prediction data, a charging and discharging strategy is dynamically generated based on a DQN algorithm; and dispatching power grid electric energy to be supplied to an electric vehicle battery or dispatching electric vehicle battery electric energy to be supplied to a power grid according to the charging and discharging strategy. The method can control the charging and discharging of the electric vehicle based on the load prediction and optimization scheduling model, balance the power grid load fluctuation, and improve the power resource utilization rate.

Description

Technical Field

[0001] The present application relates to the field of smart grids, and in particular to a method, system, device and storage medium for intelligent dispatching system of electric vehicle bidirectional charging grid load based on artificial intelligence. Background Art

[0002] With the rapid development of electric vehicles and the surge in charging demand, traditional power grids are facing greater load fluctuation pressure, especially during peak hours of electricity consumption. It is difficult to effectively manage a large number of charging demands, and the power grid is overloaded. High loads cause voltage changes, affecting the stability of household appliances and industrial equipment, and at the same time affecting the stability of the power grid. At the same time, electric vehicles have large-scale energy storage potential. If they can achieve two-way charging and discharging interaction with the power grid, it can not only reduce the fluctuation burden of the power grid, but also bring new benefits to electric vehicle users. Existing load management methods lack intelligent prediction capabilities and are difficult to dispatch the charging and discharging operations of electric vehicles in real time. Therefore, an innovative system that integrates artificial intelligence load prediction, two-way charging and discharging management, and intelligent scheduling is urgently needed. Summary of the invention

[0003] The present application discloses an intelligent dispatching method and system for electric vehicle bidirectional charging grid load based on artificial intelligence, which can control the charging and discharging of electric vehicles, balance grid load fluctuations, and improve the utilization rate of power resources based on load prediction and optimization dispatching models.

[0004] In a first aspect, the present invention provides an artificial intelligence-based intelligent dispatching method for electric vehicle bidirectional charging grid load, comprising: Regularly collect power grid load data within a preset time period in a preset power consumption area, where the power grid load data includes first load data, electric vehicle charging demand data, user behavior data, and weather data; Perform data cleaning, feature extraction, and data fusion on power grid load data to generate fused data groups; The first algorithm model processes the fused data group to obtain load forecast data of the power grid; Dynamically generate charging and discharging strategies based on grid load forecast data and the DQN algorithm; According to the charging and discharging strategy, the grid electricity is dispatched to supply the electric vehicle battery, or the electric vehicle battery electricity is dispatched to supply the grid.

[0005] In some embodiments, the first load data includes voltage data, current data, power data, and power usage period data.

[0006] In some embodiments, the power data includes active power data, reactive power data, and power factor data.

[0007] In some embodiments, weather data includes temperature data, humidity data, wind speed data, rainfall data, snowfall data, lightning data, sunshine intensity data, haze data, and dust data.

[0008] In some embodiments, user basic information, the user basic information includes user quantity data, user distribution data, user type data, and user travel habit data; Vehicle information, including vehicle ownership data, vehicle battery parameter data, and vehicle mileage data; Charging behavior information, charging behavior information charging time data, charging location preference data, charging frequency data; Market and policy data: Market and policy data include market growth rate data and policy orientation data. Market growth rate data include load scale change data and load spatiotemporal distribution change data. Policy orientation data include data on changes in user charging strategies caused by electricity price policies.

[0009] In some embodiments, the user behavior data includes timestamp data, time period data, instantaneous power data, average power data, device classification data, device quantity data, user category data, and geographic location data.

[0010] In some embodiments, data cleaning, feature extraction, and data fusion of power grid load data to generate data cleaning in the fused data set include: supplementing the missing power grid load data by mean interpolation or linear interpolation, calculating the standard deviation and mean of the corresponding power grid load data, treating data points beyond the range of ±2 times the standard deviation of the mean as outliers, and adjusting the outliers by analyzing the trend of adjacent normal data points, and identifying and deleting these duplicate values ​​by comparing the fields of the data records; The power grid load data is cleaned, feature extracted, and data fused to generate feature extraction in the fused data group, including: calculating at least one of the maximum value, minimum value, average value, and standard deviation of the power grid load data, calculating the average value of the power grid load data in a first preset time period to obtain moving average feature data, calculating the moving autocorrelation coefficient feature data packet of the first power grid load data packet at the starting time point and the second power grid load data packet at the end time point in a preset second time period, performing a fast Fourier transform on the power grid load data to obtain the Fourier transform feature value, and obtaining the rate of change data packet based on the power grid load data base calculation; Data cleaning, feature extraction, and data fusion of power grid load data to generate fused data groups include: The grid load data, charging demand data, user behavior data, and weather data are integrated to form a fused data set.

[0011] In some embodiments, the first algorithm model processes the fused data group to obtain load forecast data for the power grid, including: Convert the fused data set into matrix data, normalize the matrix data, and divide the data into training set, validation set, and test set; The first algorithm model includes an LSTM model and a GRU model. The training set is input into the LSTM model and the GRU model for pre-training. Based on their independent loss functions and optimizers, the first hidden state set and the second hidden state set are obtained respectively. The first hidden state set is fused with the second hidden state set to obtain a fused hidden state set. After the fused state set is input into the shared output layer, the power grid load forecast data is obtained. It should be noted that LSTM is a long short-term memory network model, and the GRU model is a gated recurrent unit, and the English name is Gated Recurrent Unit.

[0012] In some embodiments, dynamically generating a charging and discharging strategy based on the DQN algorithm according to the grid load forecast data includes: The environmental model is constructed by using the previous electricity price, remaining battery power, number and distribution of electric vehicles connected, and grid carrying capacity in different regions. The state vector S is expressed as: S=[L t , L t+1 ,⋯ ,L t+n , P t , E r1 , E r2 ,⋯ ,E rm , N ev , C a1 ,⋯ ,C ak ] L t Indicates the current grid load, L t+i represents the forecast load for the next i periods, where i is any value from 1 to n, P t represents the current electricity price, E rj represents the remaining power of the jth electric car, where j is any value from 1 to m, and N ev represents the total number of electric vehicles connected to the grid, C an represents the grid capacity-related index of the nth region, where n is any value from 1 to k; Define a discrete action set A, which includes at least one of fast charging, slow charging, pause charging, and battery discharge back to the grid. Input the state vector S into the neural network and output the selection probability distribution π(a|s) of each action, π probability distribution, a represents action, s represents state; The comprehensive reward function is obtained based on the grid peak-valley difference adjustment reward factor, the charging and discharging strategy economic benefit reward factor, the electric vehicle travel demand guarantee reward factor, and each action. The state vector S is updated and input into the mature DQN algorithm model. The action a is selected according to the network output action probability distribution π(a|s) and the ε-greedy strategy to regulate the charging and discharging behavior of electric vehicles.

[0013] In some embodiments, according to the grid load forecast data and based on the DQN algorithm model, the charging and discharging strategy is dynamically generated, and the mature DQN algorithm model acquisition step includes: The environmental model is constructed by using the previous electricity price, remaining battery power, number and distribution of electric vehicles connected, and grid carrying capacity in different regions. The state vector S is expressed as: S=[L t , L t+1 ,⋯ ,L t+n , P t , E r1 , E r2 ,⋯ ,E rm , N ev , C a1 ,⋯ ,C ak ] L t Represents the current grid load, which is assigned the grid load of the first historical period, L t+i represents the forecast load of the future i periods, and is assigned the forecast load of the i-th historical period, where the value of i is any one from 1 to n, P t represents the current electricity price, which is assigned the electricity price of the first historical period, E rj represents the remaining power of the j-th electric vehicle, and is assigned the remaining power value of the j-th electric vehicle in the first historical period, where the value of j is any one from 1 to m, and N ev represents the total number of electric vehicles connected to the grid, and is assigned the value of the total number of electric vehicles connected to the grid in the first historical period, C al represents the grid capacity related index of the lth region, C an represents the grid capacity-related index of the lth region, where the value of n is any one from 1 to k; Define a discrete action set A, which includes at least one of fast charging, slow charging, pause charging, and battery discharge back to the grid. Input the state vector S into the neural network and output the selection probability distribution π(a|s) of each action, π probability distribution, a represents action, s represents state; The comprehensive reward function is obtained based on the power grid peak-to-valley difference adjustment reward factor in the first historical period, the charging and discharging strategy economic benefit reward factor, the electric vehicle travel demand guarantee reward factor, and each action. Input multiple groups of state vectors S into the DQN algorithm model for training to obtain a mature DQN algorithm model. It should be noted that the English name of the DQN algorithm model is Deep Q-Network, which is a reinforcement learning algorithm model based on deep learning.

[0014] In some embodiments, according to the grid peak-valley difference adjustment reward factor, the charging and discharging strategy economic benefit reward factor, the user satisfaction reward factor and each action corresponding to the comprehensive reward function, The power grid peak-valley difference regulation reward factor includes the load peak-valley difference regulation reward factor and the overload prevention penalty factor. The load peak-valley difference regulation reward factor is expressed as: r grid,tab =k1×(|L peak −L valley ∣ prev −∣L peak −L valley ∣ new ), k1 is the weight coefficient of the load peak-valley difference adjustment reward, which is used to adjust the size of the reward. peak Indicates the peak load of the power grid, that is, the load of the power grid during peak hours, L valley Indicates the valley load of the power grid, that is, the load of the power grid during the low valley period, |L peak −L valley ∣ prev Indicates the difference between the peak and valley loads of the power grid before the action, |L peak −L valley ∣ new Indicates the difference between peak and valley loads of the power grid after the action; The overload prevention penalty factor is expressed as: r grid, verload =−k2×(L current −L safemax ), k2 is the penalty coefficient, L current为 Current load, L safemax Safe bearing upper limit load; The economic benefit reward factor of the charging and discharging strategy is expressed as: r eco =k3×(E charged ×(P high −P low ) eff ), k3 is the economic weight, E charged is the effective charge and discharge capacity (charging is positive, discharging is negative), (Phigh −P low ) eff is the effective electricity price difference after considering the loss of charging and discharging efficiency, P high represents the electricity price during peak hours, P low Indicates the electricity price during off-peak hours; The user satisfaction reward factor is expressed as: r use r=k4×(E rtarget −|E rtarget −E rj |), k4 is the user satisfaction weight, which is used to adjust the size of the reward, E rtarget is the target remaining power, i.e. the power required estimated by the user based on his / her daily travel habits, E rj is the actual remaining power, that is, the actual power of the electric vehicle at present; The comprehensive reward function can be expressed as: R(s,a)=ω1×r grid,tab +ω2×r grid,overload +ω3×r eco +ω4×r user , where ω1, ω2, ω3, and ω4 are the weights of each dimension.

[0015] In a second aspect, the present invention provides an intelligent dispatching system for electric vehicle bidirectional charging grid load based on artificial intelligence, comprising: The data collection and processing module is used to regularly collect the grid load data of the preset power consumption area, and perform data cleaning, feature extraction, and data fusion to generate a fusion data group. The grid load data includes the first load data, electric vehicle charging demand data, user behavior data, and weather data; A load forecasting module processes the fused data group through a first algorithm model to obtain load forecasting data of the power grid; Intelligent dispatch optimization module dynamically generates charging and discharging strategies based on grid load forecast data and DQN algorithm; The bidirectional charging management module is used to dispatch the grid power to replenish the electric vehicle battery according to the charging and discharging strategy, or dispatch the electric vehicle battery power to replenish the grid.

[0016] In some embodiments, the data acquisition and processing module includes: Data acquisition module, used to regularly collect and transmit power grid load data; A data storage module, used for receiving and storing the power grid load data collected by the data collection module; The data calling module is used to call the data in the data storage module; The data processing module is used to receive the power grid load data called by the first data calling module, and perform data cleaning, feature extraction, and data fusion on the power grid load data to generate a fused data group.

[0017] In a third aspect, the present invention provides an electronic device, the electronic device comprising: at least one processor; and, a memory communicatively connected to at least one processor; wherein, The memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor so that the at least one processor can execute any embodiment of the method for intelligent scheduling of electric vehicle bidirectional charging grid load based on artificial intelligence in the first aspect of the embodiment.

[0018] In a fourth aspect, the present invention provides a computer-readable storage medium storing at least one instruction, which, when executed by a processor in an electronic device, implements an embodiment of an intelligent scheduling method for loads of a bidirectional charging grid for electric vehicles based on artificial intelligence as claimed in any one of the claims.

[0019] Used to optimize grid load management, balance peak and trough electricity consumption, and increase the participation of electric vehicles in the grid.

[0020] Compared with the prior art, the beneficial effects of the present application are as follows: the intelligent dispatching method for electric vehicle bidirectional charging grid load based on artificial intelligence can integrate various types of data, including load data, real-time load data, electric vehicle charging demand data, user behavior data, and weather data, and through data cleaning, feature extraction, and data fusion technologies, it can dig out richer and more valuable feature information, thereby more comprehensively reflecting the changing laws and influencing factors of the grid load; comprehensively considering the influence of load data, real-time load data, electric vehicle charging demand data, user behavior data, and weather data factors on the grid load, it can greatly improve the accuracy of grid load data prediction, and can formulate more personalized and refined power supply strategies according to different users and different scenarios to better meet user needs, while improving the economy and reliability of grid operation. During the peak period of grid load, the charging and discharging status of electric vehicles can be adjusted in time to reduce the load impact on the grid; when a grid failure or emergency occurs, the energy storage function of the electric vehicle battery can be used to provide temporary support for the grid, thereby improving the emergency response capability and power supply reliability of the grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 It is a logic block diagram of an intelligent dispatching method for electric vehicle bidirectional charging grid load based on artificial intelligence; Figure 2 It is a schematic diagram of the intelligent dispatching system of electric vehicle bidirectional charging grid load based on artificial intelligence; Figure 3 This is another schematic diagram of the intelligent dispatching system for electric vehicle bidirectional charging grid load based on artificial intelligence; Figure 4 It is an electronic equipment schematic diagram; DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0024] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present application and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0025] It can be understood that the terms "first", "second", etc. used in the present application can be used in this article to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from another element.

[0026] The embodiments of the present application disclose an artificial intelligence-based intelligent dispatching method, system, device, and storage medium for electric vehicle bidirectional charging grid load, which are described in detail below.

[0027] refer to Figure 2, an intelligent dispatching system for electric vehicle bidirectional charging grid load based on artificial intelligence includes a data acquisition and processing module 1, a load prediction module 2, an intelligent dispatching and optimization module 3 and a bidirectional charging management module 4. The data acquisition and processing module is used to regularly collect grid load data, and perform data cleaning, feature extraction, and data fusion on the grid load data to generate a fusion data group, first load data, electric vehicle charging demand data, user behavior data, and weather data; the load prediction module 2 processes the fusion data group through a first algorithm model to obtain grid load prediction data; the intelligent dispatching and optimization module dynamically generates a charging and discharging strategy based on the grid load prediction data and the DQN algorithm; the bidirectional charging management module 4 is used to dispatch the grid power to replenish the electric vehicle battery according to the charging and discharging strategy, or dispatch the electric vehicle battery power to replenish the grid.

[0028] refer to Figure 3 In some embodiments, in an intelligent dispatching system for electric vehicle bidirectional charging grid load based on artificial intelligence, the data acquisition and processing module 1 includes a data acquisition module 11, a data storage module 12, a first data calling module 13 and a data processing module 14. The data acquisition module 11 is used to regularly collect and transmit grid load data; the data storage module 12 is used to receive and store the grid load data collected by the data acquisition module 11; the first data calling module 13 is used to call the data in the data storage module 12; the data processing module 14 is used to receive the grid load data called by the first data calling module, and perform data cleaning, feature extraction, and data fusion on the grid load data to generate a fused data group.

[0029] refer to Figure 1 In some embodiments, the method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence can be implemented based on the intelligent dispatching system for electric vehicle bidirectional charging grid load. The method is as follows: S100: regularly collecting grid load data for a preset time period in a preset power consumption area, where the grid load data includes load data, electric vehicle charging demand data, user behavior data, and weather data; S200: performing data cleaning, feature extraction, and data fusion on power grid load data to generate a fused data group; S300: The first algorithm model processes the fused data group to obtain load forecast data of the power grid; S400: Dynamically generates charging and discharging strategies based on grid load forecast data and the DQN algorithm; S500: Dispatching the grid power to supply the electric vehicle battery according to the charging and discharging strategy, or dispatching the electric vehicle battery power to supply the grid.

[0030] The load data includes, but is not limited to, voltage data, current data, power data, and power consumption period data.

[0031] The preset power consumption area is a specific power supply geographical division pre-defined under the operation and planning of the power system. The preset time period refers to a specific time interval for collecting power grid load data that is pre-set based on multiple factors such as system design objectives, power grid operation characteristics, user power consumption habits, and electric vehicle charging needs before the entire intelligent dispatching system is put into operation. This time interval can be fixed. For example, data is collected from 9 am to 9 pm every day with a collection cycle of every 10-15 minutes. During this time range, the changes in power grid load are more complex and electric vehicle charging activities are more frequent, so sufficient and rich information can be obtained for subsequent analysis and decision-making.

[0032] Specifically, since the electric energy will be distributed to various distribution network lines and user terminals after being output from the substation, voltage transformers are installed at the outlet lines of each voltage level of the substation in the preset power consumption area to monitor the voltage and obtain voltage data, current transformers are installed to detect current data, or high-precision power meters are installed to obtain power data. These voltage data can reflect the grid voltage level in the area that the substation is responsible for.

[0033] The distribution network further distributes the electric energy to each user area. Installing voltage transformers at the section switch position of the distribution network line can monitor the voltage conditions of different lines in the preset power consumption area, so as to promptly detect voltage drop problems on the line, especially in the case of long-distance power distribution or uneven load distribution, and locate the specific line section where the voltage may be too low or too high. At the same time, installing current transformers and high-precision power detectors at the section switch position of the distribution network line can obtain the corresponding current data and power data.

[0034] Voltage transformers, current transformers, and high-precision power meters can also be installed at the sectionalizing switch positions of the distribution network lines to monitor the voltage, current, and power of different lines, which can be used as an important basis for the formulation of distribution strategies, help load balancing, guide voltage regulation, influence the access control and optimization of distributed power sources, and ensure the smooth operation of the distribution system in the preset power consumption area. Voltage detection, current detection, and power detection at the user-side meter can more accurately measure the actual voltage, actual current, and actual power received by the user.

[0035] The data of electricity consumption period can be obtained through smart meters, which divide a day into multiple periods. For example, during peak hours, normal hours, and valley hours, the electricity consumption of users in each period is measured according to the pre-set period division rules. These period divisions can be customized according to local electricity price policies and grid operation characteristics. The clock module inside the meter ensures the accuracy of time, and at the end of each period, it automatically records the electricity consumption of the period and stores this data with a timestamp in the storage unit of the meter. Then, through the communication interface, these time-stamped electricity consumption data will be regularly uploaded to the data management system of the power company for analyzing the user's electricity consumption behavior pattern, formulating time-of-use electricity price strategies, and grid load forecasting. The time-of-use electricity price policy will guide users to change their charging behavior. If a valley electricity price is set, many users will adjust their charging time to the valley period, which will increase the load of the grid during this period, while the load during the peak period may decrease. This requires re-evaluating the load situation in different periods during load forecasting in order to arrange power generation plans and grid dispatching more accurately.

[0036] It should be noted that by regularly collecting grid load data including load, electric vehicle charging demand, user behavior, weather, etc. in a specific time period of a preset power consumption area, the predicted data is obtained by processing with the first algorithm model after data cleaning, feature extraction and fusion, and then the charging and discharging strategy is generated based on the DQN algorithm to achieve scheduling. It can comprehensively consider the influencing factors based on multi-source data, more accurately grasp the load change trend, and provide a reliable basis for subsequent decision-making; it can scientifically allocate electricity between the power grid and the electric vehicle battery according to the predicted and generated strategies, realize peak shaving and valley filling, optimize resource utilization, ensure the stable operation of the power grid, and enhance the ability of the power grid to cope with complex power consumption conditions and electric vehicle charging needs, and enhance the economy and safety of the overall operation.

[0037] In some embodiments, in an artificial intelligence-based intelligent dispatching method for electric vehicle bidirectional charging grid load, the power data includes active power data, reactive power data, and power factor data.

[0038] Specifically, active power is calculated by measuring the effective values ​​of voltage and current, reflecting the actual energy consumption; reactive power is related to inductance or capacitance components, affecting the voltage level; power factor is the cosine value of the phase difference between voltage and current, measuring the efficiency of the power system.

[0039] It should be noted that in the intelligent dispatching method of electric vehicle bidirectional charging grid load based on artificial intelligence, active power data reflects actual power consumption, reactive power data is related to the impact of inductance or capacitance components on voltage level, and power factor data measures the efficiency of the power system. These power data help to accurately predict the load of the power grid. Accurate prediction can know the load change trend in advance, provide a basis for load dispatch, enable the power grid to reasonably arrange power generation plans according to active and reactive power demand in different periods, optimize the proportion of traditional energy and new energy power generation, allocate resources before the peak load to ensure stable power supply, and improve power quality and reduce losses by adjusting the power factor. During load dispatching, the power grid operation status can be monitored in real time based on power data, abnormalities can be discovered and adjusted in time, peak shaving and valley filling can be achieved, the economy, safety and stability of power grid operation can be improved, and the power interaction between electric vehicles and the power grid can be effectively coordinated.

[0040] In some embodiments, the weather data includes at least one of temperature data, humidity data, wind speed data, rainfall data, snowfall data, lightning data, sunshine intensity data, haze data, and dust data.

[0041] Specifically, temperature data includes but is not limited to the daily maximum temperature, daily minimum temperature, average temperature, and temperature change rate. Temperature data may affect the use of refrigeration equipment such as air conditioners. High temperatures in summer lead to a large number of refrigeration equipment such as air conditioners, and low temperatures in winter lead to an increase in the use of heating equipment, which will increase the load on the power grid. Humidity data includes but is not limited to relative humidity and absolute humidity. Increased humidity will make people feel more stuffy, so they will use fans, air conditioners and other equipment more frequently, increasing the load on the power grid. Wind speed data includes but is not limited to average wind speed, maximum wind speed, and time distribution of wind speed. Precipitation data includes but is not limited to the size of rainfall, rainfall duration, and rainfall intensity. Snowfall data includes but is not limited to snowfall data, snowfall duration, and snow depth data. Lightning data includes but is not limited to the time, frequency, intensity, and location data of lightning. Specifically, sunshine intensity data includes but is not limited to sunshine duration, average sunshine intensity, and sunshine change data at different times. Haze data includes but is not limited to the concentration, duration, and coverage of haze. Dust data includes but is not limited to the intensity, duration, and impact range of dust.

[0042] Temperature data, humidity data, wind speed data, and rainfall data can be obtained through thermometers at meteorological observation stations, or with the help of satellite remote sensing and numerical weather forecast models.

[0043] It should be noted that applying various weather data such as temperature, humidity, wind speed, rainfall, snowfall, lightning, sunshine intensity, haze, sandstorms, etc. to grid load forecasting can consider influencing factors in multiple dimensions. For example, temperature and humidity synergistically affect the electricity consumption of refrigeration and heating equipment, wind speed and other factors are related to the power generation of renewable energy, and lightning and other factors also have their own effects, avoiding the limitation of a single factor, and capturing the complex nonlinear and dynamic relationship between weather and load, thereby improving forecast accuracy. In grid load scheduling, power generation resource allocation can be planned in advance based on accurate forecasts, and extreme weather emergencies can be flexibly responded to, promoting the consumption and coordinated complementarity of new energy, responding to extreme weather conditions, and enhancing the safety and stability of the grid, including optimizing equipment maintenance, adjusting voltage and frequency, to ensure reliable operation of the grid.

[0044] In some embodiments, the electric vehicle charging demand data includes basic user information, including user quantity data, user distribution data, user type data, and user travel habit data; vehicle information, including vehicle ownership data, vehicle battery parameter data, and vehicle range data; charging behavior information, including charging time data, charging location preference data, and charging frequency data. Market and policy data include market growth rate data and policy-oriented data, where market growth rate data includes load scale change data and load spatiotemporal distribution change data, and policy-oriented data includes relevant data on changes in user charging strategies caused by electricity price policies.

[0045] The user quantity data can determine the overall scale of electric vehicle users. The user distribution data can determine the distribution of users in geographical space. Under the preset power consumption area, the statistics can be divided according to the administrative divisions of the city, such as districts and counties; or according to functional areas such as commercial areas, residential areas, industrial areas, science and technology parks, etc.

[0046] User type data can be divided into individual users, corporate users, and operational vehicle users. There are significant differences in charging needs and behavior patterns among different types of users. The charging time of individual users is usually more flexible. They may charge at home after get off work, or use public charging piles to charge when going out for business, shopping, leisure and entertainment; corporate users may consider setting up charging piles in the company's internal parking lot to meet the charging needs of employees during work, and corporate users may have specific requirements and specifications for the management and use of charging facilities; operational vehicle users need to charge frequently to ensure the continuous operation of the vehicle due to their operational nature. Their charging time and location are relatively fixed, and they have higher requirements for charging speed and efficiency. User travel habit data includes but is not limited to the user's daily travel route, travel distance, and travel time distribution. By mining and analyzing a large amount of user travel data, we can understand the user's travel patterns.

[0047] The vehicle battery parameter data includes but is not limited to information such as battery capacity, charging rate, battery type, battery health status, etc. In some embodiments, the battery type can be a lithium-ion battery or a lithium iron phosphate battery. The battery capacity determines the maximum mileage that the vehicle can travel on a single charge, and batteries of different capacities have different charging requirements and times. The charging rate affects the charging speed, and batteries with high charging rates can charge more power in a shorter time. Determining the battery type helps determine the appropriate charging technology and charging strategy. The battery health status is also important for evaluating the changes in the vehicle's charging efficiency and cruising range. As the battery usage time increases and the number of charge and discharge times increases, the battery health status will gradually decline, which may lead to problems such as longer charging time and shorter cruising range.

[0048] The vehicle range data is closely related to the battery capacity and the energy efficiency of the vehicle, and is an important indicator for measuring how far an electric vehicle can travel on a single charge. The range data not only affects the user's travel planning and charging decisions.

[0049] Charging time data includes the specific time and duration of charging within 24 hours a day or at different time periods within a week. By analyzing the charging time data, it is possible to find obvious peak and valley periods of charging.

[0050] Charging location preference data reflects the user's inclination in choosing a charging location. Users may decide on a charging location based on their travel destination, residence, workplace, convenience of charging facilities, charging speed, charging cost, etc. The charging location can be one of the travel destination, residence, and workplace.

[0051] The charging frequency data reflects the number of times a user charges the vehicle within a certain period of time. The charging period can be one week, one month, or other periods customized according to needs. The charging frequency is related to factors such as the user's travel mileage, vehicle range, and charging habits. For example, for users with short daily travel mileage and long vehicle range, their charging frequency may be relatively low; while for users of operating vehicles, due to frequent use of their vehicles and long daily mileage, they may need to charge multiple times a day.

[0052] In some embodiments, the user behavior data includes timestamp data, time period data, instantaneous power data, average power data, device classification data, device quantity data, user category data, and geographic location data.

[0053] Specifically, timestamp data is presented in the format of year, month, day, hour, minute, and second. The time tag corresponding to a behavior clearly tracks when the user performed a certain operation, such as: the laundry home starts working at 19:30:15 on October 1, 2024, and October 1, 2024, 19:30:15 is the timestamp. Timestamp data can be used to track the user's operation sequence and the order of events, and determine the exact time when the fault occurred. Time period data The repetitive pattern of user behavior within a period of time. Instantaneous power data is the amount of electric power consumed by the electrical equipment at a specific moment. Average power data is the average value of the electric power consumed by the electrical equipment within a specific time interval. Equipment classification data includes but is not limited to equipment category, model, and function. User category data includes but is not limited to the number and scale of residential users, industrial users, commercial users, and agricultural users.

[0054] Furthermore, the market growth rate data refers to the change in the total load borne by the power grid within a preset time period in the market environment where electric vehicle charging interacts with the power grid; it can be measured by counting the total charging power of all electric vehicles connected to the power grid and the total power of other conventional power-consuming equipment in different time periods, such as hours, days, months, and years. For example, as the number of electric vehicles increases, the total load of the power grid may show a gradual upward trend; or in certain specific time periods, such as at night when the electricity price is low, a large number of electric vehicles are concentratedly charged, resulting in a significant jump in the load scale during this period. The change data of load scale is a key indicator for evaluating the market growth rate, which helps power grid companies plan the expansion and upgrade of the power grid in advance to cope with the growing demand for electricity. At the same time, it also provides a basis for power generation companies to adjust their power generation plans and reasonably arrange energy resources to ensure the stability and reliability of power supply.

[0055] Furthermore, the load temporal and spatial distribution change data refers to the distribution change of electric vehicle charging load and other power load in time and space. From the time dimension, it includes the load distribution in different periods of a day, such as peak, normal and valley load distribution, as well as seasonal load fluctuations in a week, a month or a year. From the spatial dimension, it involves the differences in load distribution in different geographical regions, such as cities, suburbs, and different power supply divisions, whose load fluctuations are different. By understanding the distribution law of load in time and space, power grid companies can formulate more accurate time-of-use electricity price strategies to guide users to charge during low load periods, thereby achieving peak shaving and valley filling of the power grid. The load temporal and spatial distribution change data includes time dimension data and space dimension data. The time dimension data can be obtained by recording the load data at each time point through smart meters and power grid data acquisition systems, and then statistically analyzing them. The space dimension data needs to be combined with the geographic information system to associate the geographical location information of each user with the load data, so as to draw a spatial distribution map of the load. At the same time, monitoring and statistics on the use of charging facilities in different regions can also provide data support for the analysis of load spatial distribution.

[0056] Furthermore, policy-oriented data refers to data related to changes in user charging strategies caused by electricity price policies. When the government or power regulatory authorities introduce electricity price policies, such as time-of-use electricity prices, peak-valley electricity prices, and tiered electricity prices, users adjust their strategies in terms of electric vehicle charging time, charging location, and charging amount. For example, when the peak-valley electricity price policy is implemented, users may transfer the charging time of electric vehicles from peak hours to valley hours in order to save electricity costs, or charge in areas with lower electricity prices. These data on changes in user charging strategies can reflect the guiding effect of policies on market behavior. By analyzing the changes in user charging strategies, we can understand whether the policy has achieved the expected goals, such as alleviating the peak load pressure of the power grid and promoting the rational use of energy. It helps them adjust their operating strategies and service content according to changes in user behavior. For example, during the valley hours when users have concentrated charging needs, the maintenance and operating costs of charging facilities can be appropriately reduced to improve service efficiency; or promotional activities and value-added services can be added at charging locations preferred by users. Policy-oriented data can record users' charging behavior data through the charging service platform, including information such as charging time, charging location, and charging amount. By comparing the changes in these data before and after the implementation of the electricity price policy, we can analyze the changes in users' charging strategies. At the same time, combined with user surveys and market feedback, we can also obtain information on users' awareness and acceptance of the electricity price policy and the reasons for changes in charging behavior, so as to more comprehensively evaluate the impact of policy guidance on users' charging strategies.

[0057] In some embodiments, performing data cleaning, feature extraction, and data fusion on power grid load data to generate a fused data set includes: S2001: Data cleaning includes supplementing the missing power grid load data by mean interpolation or linear interpolation, calculating the standard deviation and mean of the corresponding power grid load data, treating data points beyond the mean ±2 times the standard deviation as outliers, and adjusting outliers by analyzing the trend of adjacent normal data points, and identifying and deleting duplicate values ​​by comparing the fields of data records; S2002: Feature extraction includes calculating at least one of the maximum value, minimum value, average value, and standard deviation of the power grid load data, calculating the average value of the power grid load data in the first preset time period to obtain moving average feature data, calculating the moving autocorrelation coefficient feature data packet of the first power grid load data packet at the start time point and the second power grid load data packet at the end time point in the preset second time period, performing a fast Fourier transform on the power grid load data to obtain a Fourier transform feature value, and calculating the change rate data based on the power grid load data base, including: S2003: Data fusion includes integrating grid load data, charging demand data, user behavior data, and weather data to form a fused data group.

[0058] Specifically, during the data cleaning process, if the voltage data or current data is arranged into corresponding data columns in chronological order, the missing data in the data column is supplemented by the mean interpolation or linear interpolation method. Mean interpolation is to calculate the mean of the column where the data is located, and use the mean to replace the missing value and supplement it into the corresponding data column. The linear interpolation rule is to construct a linear function to estimate the missing value based on the known data points before and after the missing data. Furthermore, if the current data at a certain moment is missing, and the current value at the previous moment is I1, and the current value at the next moment is I2, and the time interval is Δt, assuming that the time from the missing moment to the previous moment is , then the missing current value I=I1+ t(I2−I1) / Δt. The same is true for the processing methods of missing other power grid load data. For missing power grid load data with linear rules, linear interpolation is used, and for missing data without linear rules, mean interpolation is selected. For the supplemented power grid load data, the corresponding standard deviation and mean are calculated. Data points beyond the range of ±2 times the standard deviation of the mean are regarded as outliers. The outliers are adjusted by analyzing the trends of adjacent normal data points, and duplicate values ​​are identified and deleted by comparing the fields of the data records.

[0059] Specifically, during the feature extraction process, the maximum value, minimum value, average value, standard deviation, etc. of the power grid load data are calculated. For example, for the voltage data within a day, the maximum value and the minimum value are found, and the average value and the standard deviation are calculated. The discrete degree and the overall level of the voltage data can be obtained. The average value of the power grid load data in the first preset time period, such as the past 1 hour, is calculated to obtain the moving average feature data.

[0060] Furthermore, suppose data is collected every m minutes, and k=60 / m data points L1, L2, ⋯, L k , then the moving average eigenvalue M A The calculation formula is: A =(L1+L2+⋯ +L k ) / k, the moving average feature can smooth short-term fluctuations and highlight long-term trends. Calculate the moving autocorrelation coefficient characteristic data packet of the first power grid load data packet at the start time point and the second power grid load data packet at the end time point in the preset second time period, such as the past 3 hours. Suppose the data packet at the start time point is P1=(x1,x2,⋯ ,x p ), the final time point data packet is P2=(y1,y2,⋯ ,y q ), x ave =(x1+x2+⋯+x p ) / p,j ave =(j1+ j2+⋯ + j p ) / p, the calculation formula of the autocorrelation coefficient r is: r = [(x1 - x ave )*(y1-y ave )+(x2- x ave )*(y2-y ave ) + ⋯ + (x p -x ave )*(y p -x ave )] / {[(x1- x ave )+(x2- x ave ) + ⋯ + (x p -x ave )]*[(y1- y ave )+(y2-y ave ) + ⋯ + (y p -x ave )]} 1 / 2 where x ave and ave are the means of P1 and P2 respectively. By calculating the autocorrelation coefficient, we can understand the correlation between data in different time periods and determine the periodicity or trend changes of the data.

[0061] Furthermore, the power grid load data is quickly Fourier transformed to obtain the Fourier transform eigenvalue. The Fourier transform can convert time domain data into frequency domain data, thereby analyzing the different frequency components in the data. For example, after the current data is Fourier transformed, the amplitude and phase information of the different frequency harmonics in the current signal can be obtained, which helps to find harmonic pollution or other frequency-related problems in the power grid.

[0062] Furthermore, the change rate data packet is calculated based on the power grid load data. For two consecutive data points D1 and D2, the calculation formula of the change rate CR is: CR = (D2-D1) / D1. The change rate feature can reflect the speed of change of the power grid load data. Each set of data in the power grid load data can be brought into the calculation one by one to obtain the corresponding change rate data packet.

[0063] In some embodiments, the first algorithm model processes the fused data group to obtain load forecast data for the power grid, including: S3001: converting the fused data set into matrix data, normalizing the matrix data, and dividing the data into a training set, a validation set, and a test set; S3002: The first algorithm model includes an LSTM model and a GRU model. The training set is input into the LSTM model and the GRU model for pre-training respectively. Based on their independent loss functions and optimizers, the first hidden state set and the second hidden state set are obtained respectively. The first hidden state is fused with the second hidden state set to obtain a fused hidden state set. After the fused state set is input into the shared output layer, the power grid load forecast data is obtained.

[0064] Converting the fused data set into matrix data enables the model to process multi-factor associations at the same time, which is conducive to exploring the comprehensive impact on the load. In terms of power grid strategy, it can comprehensively divide the electricity price period based on multiple factors; normalization allows the model to treat each feature fairly, improve the load forecasting accuracy, and provide a reliable basis for power grid dispatching strategy; reasonable data division ensures effective training and evaluation of the model, and its accurate prediction helps to accurately formulate strategies such as power generation plans. In step S3002, the LSTM model captures the long-term dependency of the load with the gating mechanism, and the GRU model has a simple structure and can quickly learn short-term features. The pre-training of the two is of great significance in long-term planning and short-term dispatching strategy formulation respectively; the hidden state set fusion combines the advantages of the two, and can more comprehensively describe the change rules in load forecasting, thereby providing support for refined management strategies such as the interaction between the power grid and electric vehicles, effectively improving the accuracy of power grid strategies and load forecasting. Through the joint training of the LSTM model and the GRU model, relevant data can be screened out to the maximum extent, so as to make more accurate predictions of the power grid load in the future, thereby improving the accuracy of power grid scheduling, thereby improving the economy, reliability and adaptability of power grid operation; at the same time, the use of a common output layer can reduce the data processing path and improve data processing efficiency.

[0065] In some embodiments, dynamically generating a charging and discharging strategy based on the DQN algorithm according to the grid load forecast data includes: S3001: The environmental model is constructed by using the previous electricity price, the remaining battery power, the number and distribution of electric vehicles connected, the grid carrying capacity of different regions and other information. The state vector S is expressed as: S=[L t , L t+1 ,⋯ ,L t+n , P t , E r1 , E r2 ,⋯ ,E rm , N ev , C a1 ,⋯ ,C ak ] L t Indicates the current grid load, L t+i represents the forecast load for the next i periods, P t represents the current electricity price, E rj represents the remaining power of the jth electric car, N ev represents the total number of electric vehicles connected to the grid, C al represents the grid capacity related indicators of the lth region; S3002: Define a discrete action set A, which includes at least one of fast charging, slow charging, suspended charging, and battery discharge back to the grid. Input the state vector S into the neural network and output the selection probability distribution π(a|s) of each action, π probability distribution, where a represents action and s represents state. S3003: Obtain a comprehensive reward function according to the power grid peak-valley difference adjustment reward factor, the charging and discharging strategy economic benefit reward factor, the electric vehicle travel demand guarantee reward factor, and each action; S3004: Update the state vector S in real time and input it into the DQN network; S3005: Update the state vector S in real time and input it into the maturely trained DQN network. Select action a according to the network output action selection probability distribution π(a|s) and ε-greedy strategy to regulate the charging and discharging behavior of electric vehicles.

[0066] In some embodiments, the step is based on the grid peak-to-valley difference adjustment reward factor, the charging and discharging strategy economic benefit reward factor, the user satisfaction reward factor, and each action corresponding to obtain a comprehensive reward function. The power grid peak-valley difference regulation reward factor includes the load peak-valley difference regulation reward factor and the overload prevention penalty factor. The load peak-valley difference regulation reward factor is expressed as: r grid,tab =k1×(|L peak −L valley ∣ prev −∣L peak −L valley ∣ new ), k1 is the weight coefficient of the load peak-valley difference adjustment reward, which is used to adjust the size of the reward. peak Indicates the peak load of the power grid, that is, the load of the power grid during peak hours, L valley Indicates the valley load of the power grid, that is, the load of the power grid during the low valley period, |L peak −L valley ∣ prev Indicates the difference between the peak and valley loads of the power grid before the action, |L peak −L valley ∣ new Indicates the difference between peak and valley loads of the power grid after the action; The overload prevention penalty factor is expressed as: r grid, verload =−k2×(L current −L safemax ), k2 is the penalty coefficient, L current为 Current load, L safemax Safe bearing upper limit load; The economic benefit reward factor of the charging and discharging strategy is expressed as: r eco =k3×(E charged ×(P high −P low ) eff ), k3 is the economic weight, E charged is the effective charge and discharge capacity (charging is positive, discharging is negative), (P high −P low ) eff is the effective electricity price difference after considering the loss of charging and discharging efficiency, P hig Represents the electricity price during peak hours. During the time of the day when electricity demand is highest, the electricity price will be relatively high. low Represents the off-peak electricity price. During the period of the day when electricity demand is the lowest, the electricity price will be relatively low; The user satisfaction reward factor is expressed as: r use r=k4×(E rtarget −|E rtarget −E rj |), k4 is the user satisfaction weight, which is used to adjust the size of the reward, E rtarget is the target remaining power, i.e. the power required estimated by the user based on his / her daily travel habits, E rj is the actual remaining power, that is, the actual power of the electric vehicle at present; The comprehensive reward function can be expressed as: R(s,a)=ω1×r grid,tab +ω2×r grid,overload +ω3×r eco +ω4×r user , where ω1, ω2, ω3, and ω4 are the weights of each dimension.

[0067] In some embodiments, the method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence dynamically generates charging and discharging strategies according to grid load forecast data and based on the DQN algorithm model, and the mature DQN algorithm model acquisition steps include: The environmental model is constructed by using the previous electricity price, remaining battery power, number and distribution of electric vehicles connected, and grid carrying capacity in different regions. The state vector S is expressed as: S=[L t , L t+1 ,⋯ ,L t+n , P t , E r1 , E r2 ,⋯ ,E rm , N ev , C a1,⋯ ,C ak ] L t Represents the current grid load, which is assigned the grid load of the first historical period, L t+i represents the forecast load of the future i periods, and is assigned the forecast load of the i-th historical period, where the value of i is any one from 1 to n, P t represents the current electricity price, which is assigned the electricity price of the first historical period, E rj represents the remaining power of the j-th electric vehicle, and is assigned the remaining power value of the j-th electric vehicle in the first historical period, where the value of j is any one from 1 to m, and N ev represents the total number of electric vehicles connected to the grid, and is assigned the value of the total number of electric vehicles connected to the grid in the first historical period, C al represents the grid capacity related index of the lth region, C an represents the grid capacity-related index of the lth region, where the value of n is any one from 1 to k; Define a discrete action set A, which includes at least one of fast charging, slow charging, pause charging, and battery discharge back to the grid. Input the state vector S into the neural network and output the selection probability distribution π(a|s) of each action, π probability distribution, a represents action, s represents state; The comprehensive reward function is obtained based on the power grid peak-to-valley difference adjustment reward factor in the first historical period, the charging and discharging strategy economic benefit reward factor, the electric vehicle travel demand guarantee reward factor, and each action. Multiple groups of state vectors S are input into the DQN algorithm model for training to obtain a mature DQN algorithm model.

[0068] It should be noted that the state vector S is formed by constructing an environmental model containing multiple information such as current and future grid load, electricity price, and remaining power of electric vehicles, and the related discrete action set A is defined. The parameters are assigned values ​​based on the data of the first historical period, and a comprehensive reward function is constructed based on multiple reward factors. The DQN algorithm model is trained to maturity using multiple sets of state vectors S. In terms of accurate grid load prediction, it can deeply mine various related information in historical data, fully consider the comprehensive impact of electric vehicle-related factors and grid characteristics on load changes, and make the prediction model more in line with actual complex scenarios, thereby improving the accuracy and reliability of predictions and accurately grasping the dynamic change trend of loads. In terms of grid load dispatching, mature DQN algorithm models can make intelligent decisions such as fast charging and discharge feedback under different grid conditions based on accurate prediction results, reasonably guide the charging and discharging behavior of electric vehicles, effectively balance the supply and demand of the grid, realize peak shaving and valley filling, optimize the allocation of grid resources, improve the flexibility and stability of the grid to cope with different load conditions, enhance the economy and safety of grid operation, ensure the efficient and orderly operation of the power system, and promote the benign interaction between electric vehicles and the grid.

[0069] refer to Figure 4 In some embodiments, the intelligent dispatching method for bidirectional charging grid load of electric vehicles based on artificial intelligence can be built into an electronic device 5, and the electronic device 5 includes at least one processor 6; and a memory 7 that is communicatively connected to the at least one processor 6; wherein the memory 7 stores instructions that can be executed by the at least one processor 6, and the instructions are executed by the at least one processor 6, so that the at least one processor 6 can execute any of the intelligent dispatching methods for bidirectional charging grid load of electric vehicles based on artificial intelligence in the present invention.

[0070] In some embodiments, a computer-readable storage medium stores at least one instruction, and when the at least one instruction is executed by a processor 6 in an electronic device 5, any of the artificial intelligence-based intelligent scheduling methods for electric vehicle bidirectional charging grid loads in the present invention is implemented.

[0071] Those skilled in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the artificial intelligence-based electric vehicle bidirectional charging grid load intelligent dispatching method and system of the present application, rather than limiting it; although the artificial intelligence-based electric vehicle bidirectional charging grid load intelligent dispatching method and system of the present application are described in detail with reference to the aforementioned embodiments, ordinary technicians in this field should understand that they can still modify the technical solutions recorded in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent dispatching method for electric vehicle bidirectional charging grid load based on artificial intelligence, characterized in that: include: Regularly collect power grid load data within a preset time period in a preset power consumption area, wherein the power grid load data includes first load data, electric vehicle charging demand data, user behavior data, and weather data; Performing data cleaning, feature extraction, and data fusion on the power grid load data to generate a fused data group; The first algorithm model processes the fused data group to obtain load forecast data of the power grid; Dynamically generate charging and discharging strategies based on grid load forecast data and the DQN algorithm; According to the charging and discharging strategy, the power of the power grid is dispatched to supply the battery of the electric vehicle, or the power of the battery of the electric vehicle is dispatched to supply the power grid.

2. According to the artificial intelligence-based intelligent dispatching method for electric vehicle bidirectional charging grid loads according to claim 1, it is characterized in that: The first load data includes voltage data, current data, power data, and power consumption period data.

3. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 2 is characterized in that: The power data includes active power data, reactive power data, and power factor data.

4. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 1 is characterized in that: The weather data includes temperature data, humidity data, wind speed data, rainfall data, snowfall data, lightning data, sunshine intensity data, haze data, and dust data.

5. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 1 is characterized in that: Electric vehicle charging demand data includes: User basic information, including user quantity data, user distribution data, user type data, and user travel habit data; Vehicle information, including vehicle ownership data, vehicle battery parameter data, and vehicle range data; Charging behavior information, including charging time data, charging location preference data, and charging frequency data; Market and policy data, including market growth rate data and policy orientation data, wherein the market growth rate data includes load scale change data and load spatiotemporal distribution change data, and the policy orientation data includes data on changes in user charging strategies caused by electricity price policies.

6. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 1 is characterized in that: The user behavior data includes timestamp data, time period data, instantaneous power data, average power data, device classification data, device quantity data, user category data, and geographic location data.

7. An artificial intelligence-based intelligent dispatching method for electric vehicle bidirectional charging grid load according to any one of claims 1 to 6, characterized in that: The data cleaning, feature extraction, and data fusion of the power grid load data to generate the fused data group includes: supplementing the missing power grid load data by mean interpolation or linear interpolation, calculating the standard deviation and mean of the corresponding power grid load data, treating the data points beyond the range of ±2 times the standard deviation of the mean as abnormal values, and adjusting the abnormal values ​​by analyzing the trend of adjacent normal data points, and identifying and deleting the duplicate values ​​by comparing the fields of the data records; The data cleaning, feature extraction, and data fusion of the power grid load data to generate feature extraction in the fused data group include: calculating at least one of the maximum value, minimum value, average value, and standard deviation of the power grid load data, calculating the average value of the power grid load data in a first preset time period to obtain moving average feature data, calculating the moving autocorrelation coefficient feature data packet of a first power grid load data packet at a starting time point and a second power grid load data packet at an end time point in a preset second time period, performing a fast Fourier transform on the power grid load data to obtain a Fourier transform feature value, and obtaining a rate of change data packet based on the power grid load data base calculation; The data fusion in the step of performing data cleaning, feature extraction, and data fusion on the power grid load data to generate a fused data group includes: The grid load data, charging demand data, user behavior data, and weather data are integrated to form a fused data set.

8. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 7 is characterized in that: The first algorithm model processes the fused data group to obtain load forecast data for the power grid, including: Converting the fused data set into matrix data, normalizing the matrix data, and dividing the data into a training set, a validation set, and a test set; The first algorithm model includes an LSTM model and a GRU model. The training set is input into the LSTM model and the GRU model for pre-training respectively. Based on their independent loss functions and optimizers, the first hidden state set and the second hidden state set are obtained respectively, and the first hidden state set is fused with the second hidden state set to obtain a fused hidden state set. After the fused state set is input into the shared output layer, the power grid load forecast data is obtained.

9. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 8 is characterized in that: The method of dynamically generating a charging and discharging strategy based on the power grid load forecast data and the DQN algorithm includes: The environmental model is constructed by using the previous electricity price, remaining battery power, number and distribution of electric vehicles connected, and grid carrying capacity in different regions. The state vector S is expressed as: S=[L t ,L t+1 ,⋯ ,L t+n ,P t ,E r1 ,E r2 ,⋯ ,E rm ,N ev ,C a1 ,⋯ ,C ak ] L t Indicates the current grid load, L t+i represents the forecast load for the next i periods, where i is any value from 1 to n, P t represents the current electricity price, E rj represents the remaining power of the jth electric car, where j is any value from 1 to m, and N ev represents the total number of electric vehicles connected to the grid, C an represents the grid capacity-related index of the nth region, where n is any value from 1 to k; Define a discrete action set A, which includes at least one of fast charging, slow charging, pause charging, and battery discharge back to the grid. Input the state vector S into the neural network and output the selection probability distribution π(a|s) of each action, π probability distribution, a represents action, s represents state; The comprehensive reward function is obtained based on the grid peak-valley difference adjustment reward factor, the charging and discharging strategy economic benefit reward factor, the electric vehicle travel demand guarantee reward factor, and each action. The state vector S is updated and input into the mature DQN algorithm model. The action a is selected according to the network output action probability distribution π(a|s) and the ε-greedy strategy to regulate the charging and discharging behavior of electric vehicles.

10. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 9, characterized in that: The step of dynamically generating a charging and discharging strategy based on the power grid load forecast data and the DQN algorithm model includes: the steps of obtaining a mature DQN algorithm model include: The environmental model is constructed by using the previous electricity price, remaining battery power, number and distribution of electric vehicles connected, and grid carrying capacity in different regions. The state vector S is expressed as: S=[L t ,L t+1 ,⋯ ,L t+n ,P t ,E r1 ,E r2 ,⋯ ,E rm ,N ev ,C a1 ,⋯ ,C ak ] L t Represents the current grid load, which is assigned the grid load of the first historical period, L t+i represents the forecast load of the future i periods, and is assigned the forecast load of the i-th historical period, where the value of i is any one from 1 to n, P t represents the current electricity price, which is assigned the electricity price of the first historical period, E rj represents the remaining power of the j-th electric vehicle, and is assigned the remaining power value of the j-th electric vehicle in the first historical period, where the value of j is any one from 1 to m, and N ev represents the total number of electric vehicles connected to the grid, and is assigned the value of the total number of electric vehicles connected to the grid in the first historical period, C al represents the grid capacity related index of the lth region, C an represents the grid capacity-related index of the lth region, where the value of n is any one from 1 to k; Define a discrete action set A, which includes at least one of fast charging, slow charging, pause charging, and battery discharge back to the grid. Input the state vector S into the neural network and output the selection probability distribution π(a|s) of each action, π probability distribution, a represents action, s represents state; The comprehensive reward function is obtained based on the power grid peak-to-valley difference adjustment reward factor in the first historical period, the charging and discharging strategy economic benefit reward factor, the electric vehicle travel demand guarantee reward factor, and each action. Multiple groups of state vectors S are input into the DQN algorithm model for training to obtain a mature DQN algorithm model.

11. The method for intelligent dispatching of electric vehicle bidirectional charging grid load based on artificial intelligence according to claim 9 is characterized in that: The step is to obtain a comprehensive reward function based on the grid peak-valley difference adjustment reward factor, the charging and discharging strategy economic benefit reward factor, the user satisfaction reward factor, and each action. The power grid peak-valley difference regulation reward factor includes the load peak-valley difference regulation reward factor and the overload prevention penalty factor. The load peak-valley difference regulation reward factor is expressed as: r grid,tab =k1×(|L peak −L valley ∣ prev −∣L peak −L valley ∣ new ), k1 is the weight coefficient of the load peak-valley difference adjustment reward, which is used to adjust the size of the reward, L peak Indicates the peak load of the power grid, that is, the load of the power grid during peak hours, L valley Indicates the valley load of the power grid, that is, the load of the power grid during the low valley period, |L peak −L valley ∣ prev Indicates the difference between the peak and valley loads of the power grid before the action, |L peak −L valley ∣ new Indicates the difference between peak and valley loads of the power grid after the action; The overload prevention penalty factor is expressed as: r grid, verload =−k2×(L current −L safemax ), k2 is the penalty coefficient, L current为 Current load, L safemax Safe bearing upper limit load; The economic benefit reward factor of the charging and discharging strategy is expressed as: r eco =k3×(E charged ×(P high −P low ) eff ), k3 is the economic weight, E charged is the effective charge and discharge capacity (charging is positive, discharging is negative), (P high −P low ) eff is the effective electricity price difference after considering the loss of charging and discharging efficiency, P hig represents the electricity price during peak hours, P low Represents the electricity price during off-peak hours; The user satisfaction reward factor is expressed as: r use r=k4×(E rtarget −|E rtarget −E rj |), k4 is the user satisfaction weight, which is used to adjust the size of the reward, E rtarget is the target remaining power, i.e. the power required estimated by the user based on his / her daily travel habits, E rj is the actual remaining power, that is, the actual power of the electric vehicle at present; The comprehensive reward function can be expressed as: R(s,a)=ω1×r grid,tab +ω2×r grid,overload +ω3×r eco +ω4×r user , where ω1, ω2, ω3, and ω4 are the weights of each dimension.

12. An intelligent dispatching system for electric vehicle bidirectional charging grid load based on artificial intelligence, characterized in that: include: A data collection and processing module is used to regularly collect power grid load data of a preset power consumption area, and to perform data cleaning, feature extraction, and data fusion on the power grid load data to generate a fusion data group, wherein the power grid load data includes first load data, electric vehicle charging demand data, user behavior data, and weather data; A load forecasting module, which processes the fused data group through a first algorithm model to obtain load forecasting data of the power grid; Intelligent dispatch optimization module dynamically generates charging and discharging strategies based on grid load forecast data and DQN algorithm; The bidirectional charging management module is used to dispatch the grid power to supply the electric vehicle battery, or dispatch the electric vehicle battery power to supply the grid according to the charging and discharging strategy.

13. The artificial intelligence-based intelligent dispatching system for electric vehicle bidirectional charging grid load according to claim 12, characterized in that: The data acquisition and processing module includes: A data acquisition module, used for regularly acquiring and transmitting the power grid load data; A data storage module, used for receiving and storing the power grid load data collected by the data collection module; A data calling module is used to call the data in the data storage module; The data processing module is used to receive the power grid load data called by the first data calling module, and perform data cleaning, feature extraction, and data fusion on the power grid load data to generate a fused data group.

14. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the artificial intelligence-based intelligent scheduling method for electric vehicle bidirectional charging grid load as described in any one of claims 1 to 11.

15. A computer-readable storage medium storing at least one instruction, characterized in that: When the at least one instruction is executed by a processor in an electronic device, an artificial intelligence-based intelligent scheduling method for electric vehicle bidirectional charging grid load is implemented as described in any one of claims 1 to 11.

Images