A method and system for predicting electric vehicle charging load and evaluating adjustable potential

Abstract

This invention discloses a method and system for predicting and assessing the charging load and adjustable potential of electric vehicles. The method uses a trip-chain-based spatiotemporal distribution model of electric vehicles to determine the spatiotemporal distribution of electric vehicles. The method analyzes the spatial coupling relationship between the electric vehicle's travel path and grid load using a road network-grid coupling model to obtain road network and traffic flow data. A speed-flow utility model is established to calculate the real-time speed of the electric vehicle. An electric vehicle power consumption model is established to analyze the electric vehicle's power consumption. Based on an electric vehicle charging demand and mode decision model, the load demand characteristics of charging stations in each region are calculated. An optimal route for the electric vehicle is planned using an optimal route planning algorithm and a travel probability matrix. Finally, the adjustable potential of the electric vehicle's charging load is assessed based on the charging demand load characteristics of each region and the optimal route for the electric vehicle. This method can accurately predict the charging load of an electric vehicle and assess its adjustable potential.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart grids and electric vehicles, and in particular relates to a method and system for predicting electric vehicle charging load and evaluating adjustable potential. Background Art

[0002] Electric vehicles (EVs), as a clean and efficient means of transportation, have been widely promoted and applied. However, their widespread adoption also poses new challenges to the stability and reliability of power grids. Due to the high randomness and volatility of EV charging loads, if not effectively predicted and managed, this can lead to grid overloads, voltage fluctuations, and other issues, impacting the normal operation of the power system.

[0003] Existing technologies for predicting EV charging loads primarily focus on the temporal distribution of charging loads, taking only a single factor into account without thoroughly considering the impact of various factors on the distribution of EV charging loads. However, EV charging loads are affected by diverse factors, and EVs exhibit temporal and spatial randomness. Different considerations will result in different charging load predictions. Charging load predictions that only consider temporal distributions can easily lead to power shortages and grid stability issues, and thus fail to meet the demands of building charging facilities and optimizing grid scheduling. Currently, the primary consideration for assessing EV scalability margin is the charging price, with little consideration given to user travel constraints. This can lead to biased estimates, hindering EV load participation in peak shaving and valley filling. Therefore, it is necessary to accurately predict EV charging loads and assess their scalability potential. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for predicting charging load and evaluating adjustable potential of electric vehicles, so as to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for predicting electric vehicle charging load and evaluating adjustable potential, comprising the following steps:

[0006] Step 1: Obtain the spatiotemporal distribution of electric vehicles through the spatiotemporal distribution model of electric vehicles based on the travel chain;

[0007] Step 2: Based on the spatiotemporal distribution of electric vehicles, the spatial coupling relationship between the driving paths of electric vehicles and the grid load is analyzed through the road network-grid coupling model to obtain road network and traffic flow data;

[0008] Step 3: Based on the road network and traffic flow data, a practical speed-flow model is established based on the real-time traffic flow statistics of the road network to calculate the real-time speed of electric vehicles;

[0009] Step 4: Based on the real-time speed and ambient temperature of the electric vehicle, establish an electric vehicle power consumption model that takes into account the ambient temperature and speed, and analyze the power consumption of the electric vehicle;

[0010] Step 5: Based on the electric vehicle charging demand and mode decision model, calculate the load demand characteristics of charging stations in each region according to the electric vehicle power consumption;

[0011] Step 6: Using the spatiotemporal distribution of electric vehicles, road network and traffic flow data, the optimal path planning algorithm and travel probability matrix are used to plan the optimal path for electric vehicles;

[0012] Step 7: Based on the charging demand load characteristics of each region and the optimal path of electric vehicles, evaluate the adjustable potential of electric vehicle charging load.

[0013] Further preferably, the steps of establishing the spatiotemporal distribution model of electric vehicles based on the travel chain are as follows:

[0014] Step 1-1: Construction of travel chain and division of urban functional areas;

[0015] Step 1-2: Determine the electric vehicle travel characteristic variables;

[0016] Steps 1-3: Fitting the electric vehicle travel characteristic variables;

[0017] Steps 1-4: Fit the first trip time in the electric vehicle travel characteristic variables;

[0018] Steps 1-5: Fit the parking time in the electric vehicle travel characteristic variables;

[0019] Steps 1-6: Fit the daily travel times in the electric vehicle travel characteristic variables;

[0020] Steps 1-7: Fit the battery capacity parameters in the electric vehicle travel characteristic variables to obtain the initial state of charge of the electric vehicle travel.

[0021] Further preferably, in step 1-1, the trip chain length refers to the number of trip segments in the trip chain, and the trip chain is divided into a simple trip chain and a complex trip chain according to the length of the trip chain; the length of a simple trip chain is 2; a complex trip chain consists of n destinations and n trip segments;

[0022] The functional areas of the city are divided into residential area H, work area W, shopping and dining area SE, social and leisure area SR, and other areas O. In order to clearly describe the travel behavior of users within a day, it is stipulated that the starting point and end point of the electric vehicle travel chain are both residential area H.

[0023] Further preferably, the road network-grid coupling model in step 2 includes a road network structure model and a grid structure model; the road network structure model G is expressed as:

[0024] ;

[0025] Where V(G) represents the set of intersection nodes, v i represents the i-th intersection node, v j represents the jth intersection node; E(G) represents the intersection node v i 、v j The road segment between the i-th intersection node and the j-th intersection node for:

[0026] ;

[0027] In the formula, inf is infinite, indicating that the road section is not connected; is the geographical length of the road segment;

[0028] is the adjacency matrix of the weighted graph corresponding to the road network topology;

[0029] Power grid structure model Expressed as:

[0030] ;

[0031] Where, is the set of grid nodes; is the number of grid nodes; is a set of power grid branches, where the elements are branches between power grid nodes; is the i-th grid node; is the jth grid node; is the power matrix of each grid node, is the active power of the ith grid node in period t, is the reactive power of the ith grid node in period t; is the coefficient matrix of load variation of each grid node; is the coefficient of change of the load of the ith grid node in period t; T is the total number of periods; is the resistance of the grid , reactance , electrical conductance and the transmission power of the branch The parameter matrix composed of .

[0032] Further preferably, the speed-flow practical model based on the real-time traffic flow statistics of the road network in step 3 is expressed as follows:

[0033] ;

[0034] in, represents the real-time speed of the electric vehicle from the i-th intersection node to the j-th intersection node at time t, Represents the zero flow velocity from the i-th intersection node to the j-th intersection node; is the maximum traffic capacity from the i-th intersection node to the j-th intersection node, which is proportional to the road grade; is the traffic flow from the i-th intersection node to the j-th intersection node at time t; and The ratio of is the road saturation at time t; β is the empirical coefficient; ζ, δ, and γ are all adaptive coefficients under different road grades.

[0035] Further preferably, the electric vehicle power consumption model in step 4 is shown as follows:

[0036] ;

[0037] ;

[0038] ;

[0039] Where, Indicates the ambient temperature is Vehicle speed Real-time power consumption per mile while driving, is the ambient temperature When the electric car is at a speed Mileage The power consumed by the air conditioner per kilometer, E is the real-time power consumption per unit mileage at different vehicle speeds, Represents the real-time speed of the electric vehicle from the i-th intersection node to the j-th intersection node; The upper limit of the air conditioner heating temperature. The lower limit of air conditioning cooling temperature; and are the cooling and heating power of the air conditioner respectively.

[0040] Further preferably, the process of establishing the decision model based on electric vehicle charging demand and mode in step 5 includes:

[0041] Step 5-1: The charging demand decision model is established as follows:

[0042] The objective factors that affect users' charging needs are shown in the following formula:

[0043] ;

[0044] Where, is the state of charge of the electric vehicle at the end of the nth trip; is the battery capacity of the electric vehicle; The real-time power consumption of the electric vehicle for the next trip; The distance of the next trip; is the state of charge safety threshold; ε is the minimum expected value of the state of charge;

[0045] Introducing remaining battery satisfaction , as shown below:

[0046] ;

[0047] Where, is the state of charge at the start of the nth trip;

[0048] The fuzzy set membership function of charging demand is shown as follows:

[0049] ;

[0050] ;

[0051] Where, Satisfaction with remaining battery The membership function of 、 are the lower and upper bounds of the charging demand fuzzy set, The ambiguity of charging requirements;

[0052] Step 5-2: The charging mode decision model is established as follows:

[0053] After the electric vehicle is fully charged, the battery state of charge is within the range [ , 1], that is, satisfying the following formula:

[0054] ;

[0055] In the formula 、 Respectively represent the charging power and charging time of electric vehicles; 、 Respectively represent the slow charging power and slow charging time; 、 They represent fast charging power and fast charging time respectively; from this we can get the relationship between charging time and charging power:

[0056] ;

[0057] Assume that the fast charge power is the slow charge power times, that is ; and let the left and right sides of the formula under the fast charging mode be 、 ,Right now:

[0058] ;

[0059] but:

[0060] ;

[0061] Therefore, the user's choice of fast charging and slow charging depends on the parking time at the destination. To judge:

[0062] ;

[0063] Further preferably, the process of calculating the load demand characteristics of charging stations in each region includes:

[0064] Step 5-3-1, calculate the state of charge of the electric vehicle at the start of each trip and the state of charge of the electric vehicle at the end of each trip;

[0065] Step 5-3-2: When an electric vehicle arrives at a destination, its charging load is aggregated to the functional area to which it belongs, and the charging load of the electric vehicles in each functional area is calculated; the charging load of each functional area is aggregated to the grid node to which it belongs, and the charging load of each grid node is obtained.

[0066] Further preferably, the optimal path planning algorithm in step 6 is an improved Floyd algorithm, which converts the path distance and the road congestion coefficient into the driving cost of the individual electric vehicle, and optimizes the total driving cost to obtain the optimal path planning between any starting and ending intersection nodes. The specific improvement steps are as follows:

[0067] Step A1: The natural process of an electric vehicle traveling from the i-th intersection to the j-th intersection is abstracted into two special states:

[0068] The first state is to use the real-time unit kilometer driving power consumption Complete path distance The total power consumed in this state is ;

[0069] The second state is the power consumption of the electric vehicle starting but stopping per unit time. Complete the waiting time for road congestion and use the road congestion coefficient Calculate the waiting time when starting but stopping during driving , the total power consumed in this state is ;

[0070] Step A2: The total amount of electricity consumed by the electric vehicle during the natural process of traveling from the i-th intersection node to the j-th intersection node is , assuming that the cost of electric vehicle individuals obtaining unit electricity from charging facilities is , then the driving cost incurred by the electric vehicle individual in the natural process of driving on the road section from the i-th intersection node to the j-th intersection node is ;

[0071] Step A3: The driving cost generated by the natural process of an electric vehicle traveling from the i-th intersection node to the j-th intersection node is ; Use driving expenses Instead of path distance Become the new impedance coefficient from the i-th intersection node to the j-th intersection node;

[0072] Step A 4: Traverse all path distances in the road network For the intersection nodes, follow steps A1 to A3 to replace and update the impedance coefficients between the corresponding intersection nodes. Substitute the updated structural impedance matrix into the Floyd algorithm to obtain the driving path with the optimal total cost between any two intersection nodes in the road network structure model.

[0073] Further preferably, the process of evaluating the adjustable potential of the electric vehicle charging load in step 7 is as follows:

[0074] Step 7-1: Establish a price difference response model, as follows:

[0075] With a step length of 1 hour, the load data of electric vehicle charging stations is taken as a sample. The load data of electric vehicle charging stations in the past six months are extracted to form a sample set. The samples in the sample set are clustered to obtain N k Initial cluster centers are generated. Each initial cluster center is used as a scenario day to draw a load change curve. The number of users in the electric vehicle charging station is determined according to the load size. Each scenario day is divided into time periods according to electricity prices. Based on the ratio of the number of users in the peak price period to the valley price period, a quantitative model for the response of electric vehicle users to the electricity price difference is established, which is expressed as:

[0076] ;

[0077] Where, represents the difference between peak and valley electricity prices, Indicates user demand responsiveness, is the maximum demand loudness, The threshold for user demand response startup is Users will participate in the electricity price demand response only when The area is called the dead zone. is the user demand response saturation threshold, that is, when the peak-valley electricity price difference is greater than When the demand responsiveness no longer increases and enters the saturation zone, and The region between is the linear region;

[0078] Step 7-2: Construct a travel demand transfer intention model based on demand response, as follows:

[0079] Based on the spatial transfer probability matrix, the next destination that electric vehicle users will reach after delaying charging due to electricity prices is predicted. Taking into account road and traffic factors, the speed-flow utility model and the electric vehicle power consumption model are combined to evaluate the minimum power required for driving at this time. Therefore, the minimum power constraint of electric vehicles when users participate in response to delayed charging due to electricity prices is obtained. Combining the minimum power constraint with the minimum power constraint that does not affect battery life, the travel demand transfer intention model is obtained as follows:

[0080] ;

[0081] ;

[0082] Where, Willingness to shift travel demand, The time period for transfer; is the battery capacity of the electric vehicle; The lowest power level that does not damage the battery life; The minimum amount of electricity required for an electric vehicle to reach its next destination. To meet the state of charge threshold required for power margin, is the state of charge safety threshold;

[0083] Step 7-3: Use the price difference transfer rate to reflect the charging time transfer probability of a single user after considering the electricity price difference during the time period, and consider the electric vehicle ξ adjustable capacity of the user responsiveness during time period t to establish an electric vehicle load adjustable potential evaluation model, as follows:

[0084] The user's transfer willingness is calculated based on the evaluation of the amount of electricity required by the electric vehicle, and the charging time of the electric vehicle is obtained from the time period. The transfer probability of the transferred electricity is:

[0085] ;

[0086] Where, Willingness to transfer for price difference; Willingness to shift travel demand;

[0087] The period The adjustable capacity of electric vehicles considering user responsiveness is:

[0088] ;

[0089] Where, For the period Electric vehicle charging capacity;

[0090] The dispatchable load capacity should be equal to the sum of the dispatchable capacities of all electric vehicles participating in the dispatch at each charging station. The evaluation model for the electric vehicle load scalability potential is:

[0091] ;

[0092] in, is the dispatchable number of electric vehicles.

[0093] The present invention also provides an electric vehicle charging load prediction and adjustable potential assessment system, comprising:

[0094] Information collection system, used to collect information data of electric vehicles, road networks, and power grids;

[0095] The data processing system is used to execute the steps of the above-mentioned electric vehicle charging load prediction and adjustable potential assessment method.

[0096] Beneficial effects of the present invention:

[0097] The system coordinates resource and demand information through a travel chain-based spatiotemporal distribution model of electric vehicles, a road network-grid coupling model, a speed-flow utility model based on real-time road network traffic statistics, an electric vehicle power consumption model, and a decision-making model based on electric vehicle charging demand and mode. Based on this, the system uses the spatiotemporal distribution of electric vehicles, road network, and traffic flow data to plan the optimal route for electric vehicles using an optimal path planning algorithm and a travel probability matrix.

[0098] The electric vehicle charging demand decision model takes into account user charging habits, charging station distribution, and the remaining power of electric vehicles to determine the load characteristics of the charging station.

[0099] An improved Floyd algorithm that integrates multi-source information is proposed to plan the optimal path of electric vehicles. The prediction takes into account comprehensive factors and uses rigorous methods, making the electric vehicle load prediction results more accurate.

[0100] By evaluating the adjustable potential of electric vehicle charging load, it is possible to more accurately provide electric vehicle aggregators with the load capacity they can participate in reducing during the grid peak regulation process, thereby improving the flexibility and security of the grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0101] Figure 1Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0102] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0103] like Figure 1 As shown, a method for predicting electric vehicle charging load and evaluating adjustable potential includes the following steps:

[0104] Step 1: Obtain the spatiotemporal distribution of electric vehicles through the spatiotemporal distribution model of electric vehicles based on the travel chain;

[0105] Step 2: Based on the spatiotemporal distribution of electric vehicles, the spatial coupling relationship between the driving paths of electric vehicles and the grid load is analyzed through the road network-grid coupling model to obtain road network and traffic flow data;

[0106] Step 3: Based on the road network and traffic flow data, a practical speed-flow model is established based on the real-time traffic flow statistics of the road network to calculate the real-time speed of electric vehicles;

[0107] Step 4: Based on the real-time speed and ambient temperature of the electric vehicle, establish an electric vehicle power consumption model that takes into account the ambient temperature and speed, and analyze the power consumption of the electric vehicle;

[0108] Step 5: Based on the electric vehicle charging demand and mode decision model, calculate the load demand characteristics of charging stations in each region according to the electric vehicle power consumption;

[0109] Step 6: Using the spatiotemporal distribution of electric vehicles, road network and traffic flow data, the optimal path planning algorithm and travel probability matrix are used to plan the optimal path for electric vehicles;

[0110] Step 7: Based on the charging demand load characteristics of each region and the optimal path of electric vehicles, evaluate the adjustable potential of electric vehicle charging load.

[0111] The distribution of electric vehicles at different time periods and spatial locations throughout the day simulates the temporal and spatial characteristics of electric vehicle travel, providing basic input data for subsequent route planning and charging demand analysis. The temporal and spatial distribution of electric vehicles includes the distribution at different time periods and spatial locations throughout the day. The electric vehicle temporal and spatial distribution model can simulate the travel time, route selection, and destination dwell time of each electric vehicle.

[0112] In this embodiment, the steps for establishing a spatiotemporal distribution model of electric vehicles based on a travel chain are as follows:

[0113] Step 1-1: Construction of travel chain and division of urban functional areas, as follows:

[0114] A trip chain describes a chain of activities in which an individual begins at a starting point, completes one or more activities, reaches different spatial locations in a specific time sequence, and finally returns to the starting point. The trip chain length refers to the number of segments in the trip chain. Trip chains can be categorized as simple or complex based on their length. A simple trip chain has a length of 2. A complex trip chain consists of n destinations and n segments, such as "home - shopping mall - ... - workspace - home." Therefore, the length of a complex trip chain is n.

[0115] According to the main functions and load types of urban areas, the functional areas of the city are divided into residential areas H, work areas W, shopping and dining areas SE, social and leisure areas SR, and other areas O. In order to clearly describe the travel behavior of users within a day, it is stipulated that the starting point and end point of the electric vehicle travel chain are both residential areas H.

[0116] Step 1-2: Determine the electric vehicle travel characteristic variables, including:

[0117] Step 1-2-1: Spatial feature variables: including the driving distance of each segment , departure point , Arrival location The spatial coordinates of the vehicle, the classification of the functional area to which it belongs, the driving path of the electric vehicle, and the node location of the distribution network where the charging location is located.

[0118] Step 1-2-2, time characteristic variables: including the departure time of each trip , end time , the duration of the trip and parking time at the destination Etc., each time characteristic quantity has the following relationship:

[0119] ;

[0120] ;

[0121] ;

[0122] Where, is the distance traveled by the electric vehicle in the nth segment; is the real-time vehicle speed from the i-th intersection node to the j-th intersection node in the n-th segment of the journey, is the departure time of the electric vehicle in the n+1th trip. It can be seen that when describing the time variable of the trip chain, it is only necessary to determine the departure time of the user's first trip. and parking duration at each destination The various time characteristic quantities of the entire travel chain can be determined.

[0123] Step 1-2-3, state characteristic variable: refers to the state of charge of the electric vehicle at the start of each journey , State of charge at the end of the trip , which is mainly affected by the user's decision-making behavior, and the current state of charge of the electric vehicle will further affect the user's next decision-making behavior.

[0124] Steps 1-3: Fit the electric vehicle travel characteristic variables. The fitting method is as follows:

[0125] In statistical principles, commonly used fitting distribution functions include mixed Gaussian distribution, Weibull distribution, rational function approximation, lognormal distribution, etc. Mixed Gaussian distribution can be simplified as:

[0126] ;

[0127] Where, is the input variable, is the output variable, 、 and are the peak value, peak position and half width of the Gaussian curve respectively. Since most of the electric vehicle travel characteristic variables are bimodal, Gaussian multi-peak fitting is used, that is, multiple Gaussian probability functions are superimposed, as shown in the following formula:

[0128] ;

[0129] Where, is the i-th peak of the Gaussian curve, is the i-th peak position of the Gaussian curve, is the i-th half width of the Gaussian curve, and M is the number of Gaussian functions. In order to judge the quality of the fitting effect, the goodness of fit test coefficient is defined as:

[0130] ;

[0131] In the formula is the goodness of fit test coefficient; N is the data dimension, is the i-th original data, For the i-th fitted regression model data , is the average value of the original data; is the sum of squares of the differences between the original data and the fitted regression model data; is the sum of the squares of the differences between the original data and its original data mean. The closer the value is to 1, the stronger the ability of the equation's variables to explain the original data, that is, the better the fitting effect.

[0132] Steps 1-4: Fit the first trip time in the electric vehicle travel characteristic variables as follows:

[0133] By performing a Gaussian fit on the first trip times of users in the trip database, we can obtain the probability densities of the four types of trips originating from residential area H. The fitting parameters are shown in the table below. It can be seen that the distribution of the first trip times for each segment of the trip chain follows a multimodal Gaussian mixture distribution composed of two unimodal Gaussian functions.

[0134]

[0135] Steps 1-5: Fit the parking duration in the electric vehicle travel characteristic variables as follows:

[0136] The parking time of electric vehicles at their destinations is determined by the nature of the destination. By fitting the distribution of the time intervals between the end of each trip and the departure time of the next trip in the travel database, we can obtain the probability distribution of the parking time of electric vehicles in functional areas other than residential area H.

[0137] The parking time distribution in work area W follows a mixed Gaussian distribution, as shown in the following formula:

[0138] ;

[0139] in is the probability density function of parking time in work zone W, and the distribution parameters are shown in the following table.

[0140]

[0141] The parking time distribution in the shopping and dining area SE follows the Weibull distribution, as shown in the following formula:

[0142] ;

[0143] In the formula The probability density function of parking time in the shopping and dining area SE is fitted with parameters a = 0.2007 and b = 1.593. The parking time in the social and leisure area SR and other areas O are all approximated by rational functions, as shown in the following equations:

[0144] ;

[0145] ;

[0146] in is the probability density function of the parking time in the social leisure area SR; the fitting parameters are p1 = 0.1549, p2 = 0.9421, q1 = -9.467, q2 = 28.5. The probability density function of parking duration in other area O is shown in the table below. The fitting parameters are p1 = 5.217, q1 = -8.306, and q2 = 36.18. The goodness-of-fit test coefficients are shown in the table below.

[0147]

[0148] Steps 1-6 are to fit the daily travel times in the electric vehicle travel characteristic variables, as follows:

[0149] According to the National Household Travel Survey (NHTS) data, the maximum number of daily trips is 9, with the average number of trips being between 1 and 4. The probability distribution is shown in the following table:

[0150]

[0151] Steps 1-7, by fitting the battery parameters in the electric vehicle travel characteristic variables, can derive the initial state of charge of the electric vehicle travel, as follows:

[0152] According to the statistical results of electric vehicle classification, the battery capacity of electric vehicles satisfies the gamma distribution shown in the following formula:

[0153] ;

[0154] Where, is the battery capacity of the electric vehicle ξ at time t; and Denotes the gamma distribution parameter. According to the electric vehicle battery data fitting results, the above parameters of the electric vehicle are set as =120kW·h, =10.08, =0.8. In order to prevent the impact of overcharging on battery safety, it is assumed that the state of charge at the initial moment is the level of the electric vehicle just after charging, and the state of charge after charging is set to 0.8-0.9 of the battery capacity. Therefore, the initial battery capacity of the electric vehicle can be obtained according to the above formula .

[0155] The electric vehicle spatiotemporal distribution model based on the travel chain of this embodiment, by introducing the concept of the travel chain, not only takes into account the travel time and spatial distribution of electric vehicles, but also combines the travel habits and destination characteristics of users. It can more accurately simulate the spatiotemporal behavior of electric vehicles and provide refined data support for grid load forecasting.

[0156] The road network-grid coupling model in step 2 of this embodiment includes a road network structure model and a grid structure model.

[0157] According to the graph theory model, the road network structure model G can be expressed as:

[0158] ;

[0159] Where V(G) represents the set of intersection nodes, v i represents the i-th intersection node, v j represents the jth intersection node; E(G) represents the intersection node v i 、v j The road segment between the i-th intersection node and the j-th intersection node for:

[0160] ;

[0161] In the formula, inf is infinite, indicating that the road section is not connected; is the geographical length of the road segment;

[0162] It is the adjacency matrix of the weighted graph corresponding to the road network topology, and its elements represent the adjacency relationship between each intersection and road section. In order to improve the accuracy of the calculation results, the longitude and latitude coordinates are used to calculate the geographical length of the road section. First, convert the latitude and longitude coordinates of the destination into a three-dimensional rectangular coordinate system. With the center of the Earth as the coordinate origin, the line connecting the center of the Earth with the equator and the 0° meridian is the x-axis, the line connecting the center of the Earth with the equator and the 90° longitude east is the y-axis, and the line connecting the center of the Earth with the North Pole is the z-axis. The relationship between the three-dimensional rectangular coordinates of the destination and its longitude and latitude is:

[0163] ;

[0164] Where, α v , β v are the longitude and latitude of a certain location v; x v 、 、 are the three-dimensional coordinates of a certain location v in the rectangular coordinate system; R e is the radius of the Earth.

[0165] According to the 3D Pythagorean Theorem, the straight-line distance (chord length) between two points can be calculated using their 3D coordinates. First, the longitude and latitude coordinates of the two points need to be converted into rectangular coordinates. Then, the straight-line distance between them can be calculated by taking the square root of the sum of the squares of the differences between these coordinates.

[0166] However, because the Earth's surface is curved, the actual surface distance (i.e., arc length) must be calculated based on the straight-line distance. The actual distance can be determined by multiplying the Earth's radius by a correction factor. This correction factor is based on the ratio of the straight-line distance to the Earth's radius, adjusted using the inverse sine function, to obtain the actual distance between the two points on the Earth's surface.

[0167] Similar to the road network, the power grid can also be simplified into nodes and branches when modeling the power grid. It can be expressed as:

[0168] ;

[0169] Where, is the set of grid nodes; is the number of grid nodes; is a set of power grid branches, where the elements are branches between power grid nodes; is the i-th grid node; is the jth grid node; is the power matrix of each grid node, is the active power of the ith grid node in period t, is the reactive power of the ith grid node in period t; is the coefficient matrix of load variation of each grid node; is the coefficient of change of the load of the ith grid node in period t; T is the total number of periods; is the resistance of the grid , reactance , electrical conductance and the transmission power of the branch The parameter matrix composed of , as shown below:

[0170] ;

[0171] Changes in road networks and traffic conditions can affect the spatial and temporal distribution of electric vehicle (EV) routes and charging demand. The road-grid coupling model considers both EV routes and grid power capacity. This model analyzes the interplay between traffic congestion and grid load, providing a basis for optimizing the layout of regional charging infrastructure.

[0172] The speed of electric vehicles directly affects the power consumption per mile, which in turn affects the charging requirements of electric vehicles. Therefore, it is necessary to accurately analyze the speed of electric vehicles at different times. In step 3 of this embodiment, a practical speed-flow model based on real-time traffic flow statistics of the road network is constructed, which is expressed as follows:

[0173] ;

[0174] in, represents the real-time speed of the electric vehicle from the i-th intersection node to the j-th intersection node at time t, Represents the zero flow velocity from the i-th intersection node to the j-th intersection node; is the maximum traffic capacity from the i-th intersection node to the j-th intersection node, which is proportional to the road grade; is the traffic flow from the i-th intersection node to the j-th intersection node at time t; and The ratio represents the road saturation at time t; β is an empirical coefficient; ζ, δ, and γ are adaptive coefficients for different road classes. This application divides the roads in the road network into two levels: ζ, δ, and γ for Level I main roads are set to 1.726, 3.15, and 3, respectively, and 2.076, 2.870, and 3 for Level II secondary roads. By simulating all electric vehicle trips at each time, the traffic flow on each road section in the road network is calculated.

[0175] The speed-flow utility model based on real-time traffic flow statistics of the road network combines real-time traffic flow and road condition data to dynamically simulate the driving speed of electric vehicles. Through the relationship model between traffic flow and speed, the driving behavior of electric vehicles and their energy demand can be more accurately predicted.

[0176] The electric vehicle power consumption model in step 4 of this embodiment is shown as follows:

[0177] ;

[0178] ;

[0179] ;

[0180] Where, Indicates the ambient temperature is Vehicle speed Real-time power consumption per mile while driving, is the ambient temperature When the electric car is at a speed Mileage The power consumed by the air conditioner per kilometer, E is the real-time power consumption per unit mileage at different vehicle speeds, Represents the real-time speed of the electric vehicle from the i-th intersection node to the j-th intersection node; The upper limit of the air conditioner heating temperature. It is the lower limit of air conditioning cooling temperature. and They are the cooling and heating power of the air conditioner, and their values ​​are 1.2kW and 1.5kW respectively.

[0181] The electric vehicle power consumption model of this embodiment not only considers the impact of the electric vehicle's driving speed on energy consumption, but also integrates ambient temperature factors, especially the impact of temperature on battery performance, so that the power consumption prediction is closer to the actual operating environment.

[0182] Furthermore, the process of calculating the load demand characteristics of charging stations in each region based on the electric vehicle charging demand and mode decision model is as follows:

[0183] Step 5-1: Establishing the charging demand decision model is as follows:

[0184] When an electric vehicle arrives at a certain location, there are two factors that affect the user's charging needs: subjective factors and objective factors. Both are closely related to whether the electric vehicle's state of charge can meet the needs of the next trip. To ensure the battery life, the electric vehicle must ensure that the battery's state of charge does not fall below the safe state of charge threshold when it arrives at the next destination. , when the electric vehicle completes the next trip and its state of charge is lower than the safety threshold, the user will charge the electric vehicle to the minimum expected state of charge value ε before the next trip, which is the objective factor affecting the user's charging demand, as shown in the following formula:

[0185] ;

[0186] Where, is the state of charge of the electric vehicle at the end of the nth trip; is the battery capacity of the electric vehicle; The real-time power consumption of the electric vehicle for the next trip; The distance of the next trip; In addition, due to the range anxiety of electric vehicle users, owners will not charge their electric vehicles until the state of charge drops to the warning value. The subjective factors of this charging demand vary from person to person and are described using fuzzy mathematics theory. In order to describe the user's satisfaction with the remaining power of the electric vehicle when it arrives at a certain destination, the remaining power satisfaction metric is introduced. , as shown below:

[0187] ;

[0188] Where, is the state of charge at the start of the nth trip;

[0189] The fuzzy set membership function of charging demand is shown as follows:

[0190] ;

[0191] ;

[0192] Where, Satisfaction with remaining battery The membership function of , whose domain is [0, 1], is the user's subjective willingness to charge; 、 are the lower and upper bounds of the charging demand fuzzy set, is the fuzziness of charging demand. When , it indicates that the remaining power is small relative to the next travel demand, or even cannot meet the next travel demand, so the user will definitely need to charge; when When the remaining power meets the needs of the next trip, but the margin is not large, and charging may be required, and The closer to l, the stronger the charging demand; when When the battery is fully charged, it indicates that the remaining power is large enough for the next trip, so the user has no intention to charge.

[0193] Step 5-2: The charging mode decision model is established as follows:

[0194] When choosing between fast and slow charging, users will carefully consider factors such as the current parking time, the amount of power required for the next trip, and battery loss. Because fast charging significantly impacts battery life, users will choose slow charging when the parking time is long enough for the next trip. When the parking time is short and slow charging is insufficient for the next trip, users will choose fast charging.

[0195] Affected by the psychological activity of range anxiety, users always have a minimum expected state of charge value ε each time they charge. Therefore, after the electric vehicle is charged, the battery state of charge should be within a certain range [ , 1], that is, satisfying the following formula:

[0196] ;

[0197] In the formula 、 Respectively represent the charging power and charging time of electric vehicles; 、 Respectively represent the slow charging power and slow charging time; 、 They represent fast charging power and fast charging time respectively. From this we can get the relationship between charging time and charging power:

[0198] ;

[0199] Assume that the fast charge power is the slow charge power times, that is ; and let the left and right sides of the formula under the fast charging mode be 、 ,Right now:

[0200] ;

[0201] but:

[0202] ;

[0203] Therefore, the user's choice of fast charging and slow charging depends on the parking time at the destination. To judge:

[0204] ;

[0205] Step 5-3: Calculate the load demand characteristics of charging stations in each region:

[0206] Step 5-3-1: The state of charge of the electric vehicle at the start of each journey is:

[0207] ;

[0208] Where, Represents whether there is a need for charging. The above formula shows that after the user makes a decision on the charging demand, if there is no need to charge, that is , then the state of charge at the start of this trip is the same as the state of charge at the end of the previous trip; if charging is required, that is , then the state of charge at the start of this trip is the state of charge of the electric vehicle after charging is completed at the destination of the previous trip.

[0209] The state of charge of the electric vehicle at the end of each journey is:

[0210] ;

[0211] Where, is the real-time power consumption of the kth road section in each trip of the electric vehicle; is the travel distance of the kth road segment; The number of road segments in the route planned by the modified Floyd algorithm for each trip.

[0212] Step 5-3-2: Charging load calculation is as follows:

[0213] When an electric vehicle arrives at a destination, its charging load is aggregated to the functional area to which it belongs. The charging load of electric vehicles in each functional area is It can be expressed as:

[0214] ;

[0215] Where, is the number of electric vehicles in functional area z during period t; is the charging power of the electric vehicle during period t.

[0216] By aggregating the charging load of each functional area to the grid node to which it belongs, the charging load of each grid node can be obtained:

[0217] ;

[0218] Where, is the total load of the g-th grid node in period t, represents that functional area z belongs to the gth grid node; is the base load of functional zone z during period t; is the charging load of functional area z during period t.

[0219] In step 6 of this embodiment, the establishment of the travel probability matrix (OD) is as follows:

[0220] The core of the travel analysis method is the trip probability matrix, whose elements represent the traffic volume between the starting and ending points of a road during that time period. By accessing observational data from the transportation department, the traffic volume of electric vehicles on a road section during each time period can be obtained, and the corresponding travel probability matrix can be calculated. Based on the observed traffic volume at each section of the road network and the seed travel probability matrix data, the trip probability matrix is ​​estimated by using one of the following allocation methods: randomized user balance method, all-or-nothing method, etc., through multiple iterations of traffic allocation and matrix estimation.

[0221] If a day is divided into 24 hours, the travel probability matrix can be seen as 24 parts, each of which consists of 1 submatrix m represents the number of nodes in the region, t=0, 1, ⋯, 23, It represents the traffic volume between the starting point and the end point of the road from time t to time t+1. is an element in the travel probability matrix A, representing the number of electric vehicles traveling from the i-th intersection node to the j-th intersection node during the period t to t+1. During the period t to t+1, the number of electric vehicles traveling from the i-th intersection node to the j-th intersection node is divided by the sum of the number of electric vehicles traveling from the i-th intersection node to any other intersection node, representing the probability that the electric vehicles starting from the i-th intersection node in the period t to t+1 take the j-th intersection node as the destination, that is:

[0222] ;

[0223] Where: is an element in the travel probability matrix; Represents the number of electric vehicles traveling from the i-th intersection node to the j-th intersection node during the period from t to t+1 ; Represents the sum of the number of electric vehicles traveling from the i-th intersection node to any other node from period t to period t+1.

[0224] The optimal path planning algorithm described in step 6 of this embodiment is the improved Floyd algorithm.

[0225] The implementation principle of the Floyd algorithm is described as follows:

[0226] Input: Impedance coefficient between any two intersection nodes in the road network structure;

[0227] Output: shortest path distance storage matrix and shortest path node storage matrix between any two intersection nodes;

[0228] Step 6-1: Initialization; Generate the structural impedance matrix D of the traffic network using the impedance coefficient between any two intersection nodes;

[0229] Step 6-2: Define the def Floyd(graph) function and initialize the path distance storage matrix and path node storage matrix with the same dimensions as the structural impedance matrix D. represents the shortest path length from the i-th intersection node to the j-th intersection node, Indicates the intermediate intersection nodes that the shortest path from the i-th intersection node to the j-th intersection node passes through;

[0230] Step 6-3: Set the starting, middle and ending intersection nodes Perform traversal loop and gradually update the path distance storage matrix and path node storage matrix according to the judgment conditions;

[0231] Step 6-4: Update the judgment condition: Then the path distance storage matrix is ​​updated as follows: ; represents the shortest path length from the i-th intersection node to the k-th intersection node, Indicates the shortest path length from the kth intersection node to the jth intersection node;

[0232] Step 6-5: Update the judgment condition: Then the path distance storage matrix is ​​updated as follows: ; Indicates the intermediate nodes passed by the shortest path from the i-th intersection node to the k-th intersection node; Indicates the intermediate intersection nodes that the shortest path from the kth intersection node to the jth intersection node passes through;

[0233] Step 6-6: Loop through all , returns the path distance storage matrix and path node storage matrix.

[0234] The traditional Floyd algorithm only minimizes the driving path and does not take the road congestion coefficient into account. Therefore, we will establish an improved Floyd algorithm to convert the path distance and road congestion coefficient into the driving cost of individual electric vehicles. The optimal path planning between any starting and ending intersection nodes is calculated by optimizing the total driving cost. The specific improvement steps are as follows:

[0235] Step A1: The natural process of an electric vehicle traveling from the i-th intersection to the j-th intersection is abstracted into two special states:

[0236] The first state is to use the real-time unit kilometer driving power consumption Complete path distance The total power consumed in this state is ;

[0237] The second state is the power consumption of the electric vehicle starting but stopping per unit time. Complete the waiting time for road congestion and use the road congestion coefficient Calculate the waiting time when starting but stopping during driving , the total power consumed in this state is ;

[0238] Step A2: The total amount of electricity consumed by the electric vehicle during the natural process of traveling from the i-th intersection node to the j-th intersection node is , assuming that the cost of electric vehicle individuals obtaining unit electricity from charging facilities is , then the driving cost incurred by the electric vehicle individual in the natural process of driving on the road section from the i-th intersection node to the j-th intersection node is ;

[0239] Step A3: The driving cost generated by the natural process of an electric vehicle traveling from the i-th intersection node to the j-th intersection node is ; Use driving expenses Instead of path distance Become the new impedance coefficient from the i-th intersection node to the j-th intersection node;

[0240] Step A 4: Traverse all path distances in the road network For the intersection nodes, follow steps A1 to A3 to replace and update the impedance coefficients between the corresponding intersection nodes. Substitute the updated structural impedance matrix into the Floyd algorithm to obtain the driving path with the optimal total cost between any two intersection nodes in the road network structure model.

[0241] The process of evaluating the adjustable potential of the electric vehicle charging load in step 7 of this embodiment is as follows:

[0242] Step 7-1: Establish a price difference response model, as follows:

[0243] With a step length of 1 hour, the load data of electric vehicle charging stations is taken as a sample. The load data of electric vehicle charging stations in the past six months are extracted to form a sample set. The samples in the sample set are clustered to obtain N k Initial cluster centers are generated. Each initial cluster center is used as a scenario day to draw a load change curve. The number of users in the electric vehicle charging station is determined according to the load size. Each scenario day is divided into time periods according to electricity prices. Based on the ratio of the number of users in the peak price period to the valley price period, a quantitative model for the response of electric vehicle users to the electricity price difference is established, which is expressed as:

[0244] ;

[0245] Where, represents the difference between peak and valley electricity prices, Indicates user demand responsiveness, is the maximum demand loudness, The threshold for user demand response startup is Users will participate in the electricity price demand response only when The area is called the dead zone. is the user demand response saturation threshold, that is, when the peak-valley electricity price difference is greater than When the demand responsiveness no longer increases and enters the saturation zone, and The area between is the linear region.

[0246] The specific process of sample clustering in the sample set is: randomly select a sample from the sample set as the first cluster initial center, calculate the remaining samples The distance from the initial center of the cluster, record the shortest distance d( ); For each sample, the probability of being selected as the next cluster center is , select the cluster center of this round according to probability, the shortest distance of the sample The larger the value, the higher the probability of being selected. Repeatedly calculate the distance between the remaining samples and the cluster center until N k Initial cluster centers.

[0247] Through the above steps, a detailed price difference response model can be established to evaluate the response behavior of electric vehicle users under different electricity price conditions. This is of great significance for optimizing power grid management and formulating operational strategies for electric vehicle charging stations.

[0248] Step 7-2: Construct a travel demand transfer intention model based on demand response, as follows:

[0249] Based on the spatial transfer probability matrix, the next destination that electric vehicle users will reach after delaying charging due to electricity prices is predicted. Taking into account road and traffic factors, the speed-flow utility model and the electric vehicle power consumption model are combined to evaluate the minimum power required for driving at this time. Therefore, the minimum power constraint of electric vehicles when users participate in response to delayed charging due to electricity prices is obtained. Combining the minimum power constraint with the minimum power constraint that does not affect battery life, the travel demand transfer intention model is obtained as follows:

[0250] ;

[0251] ;

[0252] Where, Willingness to shift travel demand, The time period for transfer; is the battery capacity of the electric vehicle; The lowest power level that does not damage the battery life; The minimum amount of electricity required for an electric vehicle to reach its next destination. To meet the state of charge threshold required for power margin, is the state of charge safety threshold.

[0253] Step 7-3: Use the price difference transfer rate to reflect the charging time transfer probability of a single user after considering the electricity price difference during the time period, and consider the electric vehicle ξ adjustable capacity of the user responsiveness during time period t to establish an electric vehicle load adjustable potential evaluation model, as follows:

[0254] The user's transfer willingness is calculated based on the evaluation of the amount of electricity required by the electric vehicle, and the charging time of the electric vehicle is obtained from the time period. The transfer probability of the transferred electricity is:

[0255] ;

[0256] Where, Willingness to transfer for price difference; willingness to shift travel demand;

[0257] The period The adjustable capacity of electric vehicles considering user responsiveness is:

[0258] ;

[0259] Where, For the period Electric vehicle charging capacity;

[0260] The dispatchable load capacity should be equal to the sum of the dispatchable capacities of all electric vehicles participating in the dispatch at each charging station. The evaluation model for the electric vehicle load scalability potential is:

[0261] ;

[0262] in, is the dispatchable number of electric vehicles.

[0263] The present invention also provides an electric vehicle charging load prediction and adjustable potential assessment system, comprising:

[0264] Information collection system, used to collect information data of electric vehicles, road networks, and power grids;

[0265] The data processing system is used to execute the steps of the above-mentioned electric vehicle charging load prediction and adjustable potential assessment method.

[0266] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for predicting electric vehicle charging load and evaluating adjustable potential, characterized in that: The following steps are involved: Step 1: Obtain the spatiotemporal distribution of electric vehicles through the spatiotemporal distribution model of electric vehicles based on the travel chain; Step 2: Based on the spatiotemporal distribution of electric vehicles, the spatial coupling relationship between the driving paths of electric vehicles and the grid load is analyzed through the road network-grid coupling model to obtain road network and traffic flow data; Step 3: Based on the road network and traffic flow data, a practical speed-flow model is established based on the real-time traffic flow statistics of the road network to calculate the real-time speed of electric vehicles; Step 4: Based on the real-time speed and ambient temperature of the electric vehicle, establish an electric vehicle power consumption model that takes into account the ambient temperature and speed, and analyze the power consumption of the electric vehicle; Step 5: Based on the electric vehicle charging demand and mode decision model, calculate the load demand characteristics of charging stations in each region according to the electric vehicle power consumption; Step 6: Using the spatiotemporal distribution of electric vehicles, road network and traffic flow data, the optimal path planning algorithm and travel probability matrix are used to plan the optimal path for electric vehicles; Step 7: Based on the charging demand load characteristics of each region and the optimal path of electric vehicles, evaluate the adjustable potential of electric vehicle charging load; The establishment of the travel probability matrix is ​​as follows: Divide a day into 24 hours, then the travel probability matrix can be seen as 24 parts, each part consists of 1 sub-matrix Composition, m represents the number of nodes in the region, t=0, 1, ⋯, 23, It represents the traffic volume between the starting point and the end point of the road from time t to time t+1. is an element in the travel probability matrix A, representing the number of electric vehicles traveling from the i-th intersection node to the j-th intersection node during the period t to t+1. During the period t to t+1, the number of electric vehicles traveling from the i-th intersection node to the j-th intersection node is divided by the sum of the number of electric vehicles traveling from the i-th intersection node to any other intersection node, representing the probability that the electric vehicles starting from the i-th intersection node take the j-th intersection node as the destination during the period t to t+1, that is: ； Where: is an element in the travel probability matrix; Represents the number of electric vehicles traveling from the i-th intersection node to the j-th intersection node during the period from t to t+1 ; represents the sum of the number of electric vehicles traveling from the i-th intersection node to any other node from period t to period t+1; The optimal path planning algorithm described in step 6 is an improved Floyd algorithm, which converts the path distance and road congestion coefficient into the driving cost of the individual electric vehicle, and optimizes the total driving cost to obtain the optimal path planning between any starting and ending intersection nodes. The specific improvement steps are as follows: Step A1: The natural process of an electric vehicle traveling from the i-th intersection to the j-th intersection is abstracted into two special states: The first state is to use the real-time unit kilometer driving power consumption Complete path distance The total power consumed in this state is ; The second state is the power consumption of the electric vehicle starting but stopping per unit time. Complete the waiting time for road congestion and use the road congestion coefficient Calculate the waiting time when starting but stopping during driving , the total power consumed in this state is ; Step A2: The total amount of electricity consumed by the electric vehicle during the natural process of traveling from the i-th intersection node to the j-th intersection node is , assuming that the cost of electric vehicle individuals obtaining unit electricity from charging facilities is , then the driving cost incurred by the electric vehicle individual in the natural process of driving on the road section from the i-th intersection node to the j-th intersection node is ; Step A3: The driving cost generated by the natural process of an electric vehicle traveling from the i-th intersection node to the j-th intersection node is ; Use driving expenses Instead of path distance Become the new impedance coefficient from the i-th intersection node to the j-th intersection node; Step A4: Traverse all path distances in the road network For the intersection nodes, follow steps A1 to A3 to replace and update the impedance coefficients between the corresponding intersection nodes. Substitute the updated structural impedance matrix into the Floyd algorithm to obtain the driving path with the optimal total cost between any two intersection nodes in the road network structure model.

2. The method for predicting and evaluating the charging load of electric vehicles and the adjustable potential according to claim 1, characterized in that: The steps to establish the spatiotemporal distribution model of electric vehicles based on the travel chain are as follows: Step 1-1: Construction of travel chain and division of urban functional areas; Step 1-2: Determine the electric vehicle travel characteristic variables; Steps 1-3: Fitting the electric vehicle travel characteristic variables; Steps 1-4: Fit the first trip time in the electric vehicle travel characteristic variables; Steps 1-5: Fit the parking time in the electric vehicle travel characteristic variables; Steps 1-6: Fit the daily travel times in the electric vehicle travel characteristic variables; Steps 1-7: Fit the battery capacity parameters in the electric vehicle travel characteristic variables to obtain the initial state of charge of the electric vehicle travel.

3. The method for predicting and evaluating the charging load of electric vehicles and the adjustable potential according to claim 2, characterized in that: In step 1-1, the trip chain length refers to the number of trip segments in the trip chain. Trip chains are divided into simple trip chains and complex trip chains based on their length. The length of a simple trip chain is 2; a complex trip chain consists of n destinations and n trip segments. The functional areas of the city are divided into residential area H, work area W, shopping and dining area SE, social and leisure area SR, and other areas O. It is stipulated that the starting point and end point of the electric vehicle travel chain are both residential area H.

4. The method for predicting and evaluating the charging load of electric vehicles and the adjustable potential according to claim 1, wherein: The road network-grid coupling model in step 2 includes a road network structure model and a grid structure model; the road network structure model G is expressed as: ； Where V(G) represents the set of intersection nodes, v i represents the i-th intersection node, v j represents the jth intersection node; E(G) represents the intersection node v i 、v j The road segment between the i-th intersection node and the j-th intersection node for: ； In the formula, inf is infinite, indicating that the road section is not connected; is the geographical length of the road segment; is the adjacency matrix of the weighted graph corresponding to the road network topology; Power grid structure model Expressed as: ； Where, is the set of grid nodes; is the number of grid nodes; is a set of power grid branches, where the elements are branches between power grid nodes; is the i-th grid node; is the jth grid node; is the power matrix of each grid node, is the active power of the ith grid node in period t, is the reactive power of the ith grid node in period t; is the coefficient matrix of load variation of each grid node; is the coefficient of change of the load of the ith grid node in period t; T is the total number of periods; is the resistance of the grid , reactance , electrical conductance and the transmission power of the branch The parameter matrix composed of .

5. The method for predicting and evaluating the charging load of electric vehicles and the adjustable potential according to claim 4, characterized in that: The speed-flow practical model based on real-time traffic flow statistics of the road network described in step 3 is expressed as follows: ； in, represents the real-time speed of the electric vehicle from the i-th intersection node to the j-th intersection node at time t, Represents the zero flow velocity from the i-th intersection node to the j-th intersection node; is the maximum traffic capacity from the i-th intersection node to the j-th intersection node, which is proportional to the road grade; is the traffic flow from the i-th intersection node to the j-th intersection node at time t; and The ratio of is the road saturation at time t; β is the empirical coefficient; ζ, δ, and γ are all adaptive coefficients under different road grades.

6. The method for predicting and evaluating the charging load of electric vehicles and the adjustable potential according to claim 5, characterized in that: The electric vehicle power consumption model described in step 4 is shown as follows: ； ； ； Where, Indicates the ambient temperature is Vehicle speed Real-time power consumption per mile while driving, is the ambient temperature When the electric car is at a speed Mileage The power consumed by the air conditioner per kilometer, E is the real-time power consumption per unit mileage at different vehicle speeds, Represents the real-time speed of the electric vehicle from the i-th intersection node to the j-th intersection node; The upper limit of the air conditioner heating temperature. The lower limit of air conditioning cooling temperature; and are the cooling and heating power of the air conditioner respectively.

7. The method for predicting and evaluating the charging load of electric vehicles and the adjustable potential according to claim 6, characterized in that: The process of establishing the decision model based on electric vehicle charging demand and mode described in step 5 includes: Step 5-1: The charging demand decision model is established as follows: The objective factors that affect users' charging needs are shown in the following formula: ； Where, is the state of charge of the electric vehicle at the end of the nth trip; is the battery capacity of the electric vehicle; The real-time power consumption of the electric vehicle for the next trip; The distance of the next trip; is the state of charge safety threshold; ε is the expected value of the minimum state of charge; Introducing remaining battery satisfaction , as shown below: ； Where, is the state of charge at the start of the nth trip; The fuzzy set membership function of charging demand is shown as follows: ； ； Where, Satisfaction with remaining battery The membership function of 、 are the lower and upper bounds of the charging demand fuzzy set, The ambiguity of charging requirements; Step 5-2: The charging mode decision model is established as follows: After the electric vehicle is fully charged, the battery state of charge is within the range [ , 1], that is, satisfying the following formula: ； In the formula 、 Respectively represent the charging power and charging time of electric vehicles; 、 Respectively represent the slow charging power and slow charging time; 、 They represent fast charging power and fast charging time respectively; thus, the relationship between charging time and charging power is obtained: ； Assume that the fast charge power is the slow charge power times, that is ; and let the left and right sides of the formula under the fast charging mode be 、 ,Right now: ； but: ； Therefore, the user's choice of fast charging and slow charging is determined by the parking time at the destination: ； Where, The duration of parking at the destination.

8. The method for predicting and evaluating the charging load of electric vehicles and the adjustable potential according to claim 1, characterized in that: The process of evaluating the adjustable potential of electric vehicle charging load described in step 7 is as follows: Step 7-1: Establish a price difference response model, as follows: With a step length of 1 hour, the load data of electric vehicle charging stations is taken as a sample. The load data of electric vehicle charging stations in the past six months are extracted to form a sample set. The samples in the sample set are clustered to obtain N k Initial cluster centers; Taking each initial cluster center as a scenario day, draw the load change curve, determine the number of users in the electric vehicle charging station according to the load size, divide the time period according to the electricity price for each scenario day, and establish a quantitative model of the response of electric vehicle users to the electricity price difference based on the ratio of the number of users in the peak electricity price period to the valley electricity price period, which is expressed as: ； Where, represents the difference between peak and valley electricity prices, Indicates user demand responsiveness, is the maximum demand loudness, The threshold for user demand response startup is Users will participate in the electricity price demand response only when The area is called the dead zone. is the user demand response saturation threshold, that is, when the peak-valley electricity price difference is greater than When the demand responsiveness no longer increases and enters the saturation zone, and The region between is the linear region; Step 7-2: Construct a travel demand transfer intention model based on demand response, as follows: Based on the spatial transfer probability matrix, the next destination that electric vehicle users will reach after delaying charging due to electricity prices is predicted. Taking into account road and traffic factors, the speed-flow utility model and the electric vehicle power consumption model are combined to evaluate the minimum power required for driving at this time. Therefore, the minimum power constraint of electric vehicles when users participate in response to delayed charging due to electricity prices is obtained. Combining the minimum power constraint with the minimum power constraint that does not affect battery life, the travel demand transfer intention model is obtained as follows: ； ； Where, Willingness to shift travel demand, The time period for transfer; is the battery capacity of the electric vehicle; The lowest power level that does not damage the battery life; The minimum amount of electricity required for an electric vehicle to reach its next destination. To meet the state of charge threshold required for power margin, is the state of charge safety threshold; Step 7-3: Use the price difference transfer rate to reflect the charging time transfer probability of a single user after considering the electricity price difference during the time period, and consider the electric vehicle ξ adjustable capacity of the user responsiveness during time period t to establish an electric vehicle load adjustable potential evaluation model, as follows: The user's transfer willingness is calculated based on the evaluation of the amount of electricity required by the electric vehicle, and the charging time of the electric vehicle is obtained from the time period. The transfer probability of the transferred electricity is: ； Where, Willingness to transfer for price difference; Willingness to shift travel demand; The period The adjustable capacity of electric vehicles considering user responsiveness is: ； Where, For the period Electric vehicle charging capacity; The dispatchable load capacity should be equal to the sum of the dispatchable capacities of all electric vehicles participating in the dispatch at each charging station. The evaluation model for the electric vehicle load scalability potential is: ； in, is the dispatchable number of electric vehicles.

9. An electric vehicle charging load prediction and adjustable potential assessment system, characterized in that: include: Information collection system, used to collect information data of electric vehicles, road networks, and power grids; A data processing system for executing the steps of the electric vehicle charging load prediction and adjustable potential assessment method as described in any one of claims 1 to 8.

Images