Active court selection method and system based on space-time transfer characteristics of electric vehicles

Abstract

The invention discloses an active court selection method and system based on time-space transfer characteristics of an electric vehicle, and mainly relates to the technical field of electric vehicle charging network optimization. Comprising the following steps: acquiring electric vehicle driving data and transformer area data, and preprocessing the acquired data to obtain the charging demand probability of the area; selecting and optimizing a zone area for charging the electric vehicle; and determining the weight in the active transformer area selection, and formulating a dynamic weight adjustment strategy. The method has the advantages that the key courts are selected and optimized according to actual electric vehicle driving and charging and discharging data, local overload can be avoided, and stability and reliability of a power grid are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle charging network optimization, and in particular to an active station area selection method and system based on the spatiotemporal transfer characteristics of large-scale electric vehicles. Background Art

[0002] With the popularization of electric vehicles, the demand for charging has increased dramatically. The random mobility of electric vehicles and the high power of electric vehicle aggregates after large-scale access will change the existing load level of the power grid, which has a great impact on the distribution network, especially the voltage stability of the substation. The large-scale disordered charging after connecting electric vehicles to charging piles may further increase the peak-to-valley difference and even affect the power supply quality, which will have a certain degree of adverse impact on safety. The random charging behavior of a large number of electric vehicles will produce a clustering effect that affects the operation of the power grid. Research on the distribution network at the substation level can help understand and predict these changes, and provide data support for future charging infrastructure planning and power grid upgrades.

[0003] Electric vehicles can be regarded as distributed energy storage devices with mobile characteristics. This new type of controllable load can transmit electricity to the distribution network through battery discharge. For a city with a large geographical area, the area is usually composed of hundreds of substations. It may be impractical to study and manage all substations. Actively selecting representative substations can simplify research and management work and focus on solving the most critical problems.

[0004] In the past, demand response of large urban power grids was mainly carried out at the city level, and it was impossible to accurately adjust the local power supply and demand. Therefore, an active area selection method and system based on the spatiotemporal transfer characteristics of electric vehicles is urgently needed to solve this problem. Summary of the invention

[0005] The purpose of the present invention is to provide an active area selection method and system based on the time-space transfer characteristics of electric vehicles. By selecting and optimizing key areas through actual electric vehicle driving and charging and discharging data, local load overload can be avoided and the stability and reliability of the power grid can be ensured.

[0006] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0007] On the one hand, a method for actively selecting a station area based on the spatiotemporal transfer characteristics of electric vehicles is provided, comprising the following steps:

[0008] S1: Collect electric vehicle driving data and substation data, and pre-process the collected data to obtain the charging demand probability of the area;

[0009] S2: Select and optimize the charging area for electric vehicles;

[0010] S3: Determine the weights in active station selection and develop a dynamic weight adjustment strategy.

[0011] Preferably, in step S1, the collected electric vehicle driving data includes: timestamp, geographical location of the electric vehicle and SoC, and the driving data is stored as time series data D i，t ={t,x i ,y i , SoC i}, where x i represents the horizontal coordinate of the i-th car at time t, y i represents the vertical coordinate of the i-th car at time t, SoC i represents the SoC of the i-th car at time t;

[0012] The obtaining of the charging demand probability of the area includes:

[0013] S11: Divide all the stations into equal proportions, and each grid area G j The size is Δx×Δy, where Δx represents the horizontal coordinate of the area division size and Δy represents the vertical coordinate of the area division size.

[0014] S12: For each grid area G j , according to the number of electric vehicles entering the area and the proportion of low-power vehicles, calculate the charging demand probability P of the area j :

[0015]

[0016] Among them, N is the total number of electric vehicles in the area, M represents the total number of all divided areas, and L j Represents the total number of charging times for the jth station.

[0017] Preferably, the step S1 further includes: based on the charging demand probability P j The initial area selection strategy is to prioritize the area with the largest charging demand G j* As the first station:

[0018] j * = arg max j P j .

[0019] Preferably, the step S2 comprises:

[0020] S21: Correct the possibility of charging electric vehicles, specifically: for each area G j , its W j represents the correction of the possibility of all electric vehicles charging in this area, W jThe calculation process is:

[0021] W j =αP j +β(1-SoC i )

[0022] Among them, α and β are coefficients for adjusting each weight;

[0023] S22: Execute the weighted K-center greedy algorithm, including:

[0024] o i The value is (x i ,y i ), (x i ,y i ) represents the longitude and latitude of the i-th electric car, O represents the original data set, m represents the sum of the points in the data set;

[0025] Construct a K nearest neighbor indicator matrix:

[0026] P=[o ij ] m×n

[0027] Among them, p ij Used to indicate o j Does it mean o i , if m j Indicates o i , then p ij =1, otherwise p ij =0;

[0028] Let p ij is replaced by the weights of the Gaussian kernel distance:

[0029]

[0030] W=[w ij ] m×n , w ij ∈[0, 1]

[0031] Among them, d x (x i ,y i ) refers to the Manhattan distance, and v represents all (x i ,y i ), and normalize the weights for a large number of K nearest neighbor edges:

[0032] Gra(O)=[Gra(o 1 ),…,Gra(o j ),…,Gra(o m )] T;

[0033] After selecting the high-density data points, the neighboring points o j Try density decay:

[0034]

[0035] Gra(o j )=Gra(o j )-Gra(o i )·p ij , o j ∈neighbour(o i )

[0036] Among them, neighbor(o i ) gives the o obtained according to the matrix P i The neighbor point set, p i is the i-th row of P, Index(p i >0) gives the value that satisfies p i > 0, based on the density of the graph, select the source with the highest density and iteratively reduce the density of neighboring sources.

[0037] Preferably, in step S22, iteratively reducing the density of neighbor sources includes:

[0038] S221: Select an initial station area as a starting point and add it to the set of marked data points;

[0039] S222: In each iteration, the coverage of the current set of marked data points is calculated, and a new station area is selected to maximize the new coverage;

[0040] S223: Add the newly selected area to the set of marked data points and update the coverage:

[0041]

[0042] Among them, e 0 is the set of currently labeled data points, e 1 is a set of selected labeled data points, δ describes e 1 For 0 The coverage radius is expressed in terms of e 0 A set of balls with a radius of δ as the center of each member can cover the entire e 1 , minimization is equivalent to the minimax station selection problem:

[0043]

[0044] Where Δ represents the calculated Euclidean distance, which is solved by using the 2-OPT solution, iterating the following equation:

[0045]

[0046] e∈e∪{o}

[0047] Where o is the data point selected in one iteration;

[0048] Steps S221 and S222 are repeatedly executed until all stations are calculated.

[0049] Preferably, in step S3, the weight in active station area selection is determined as follows:

[0050] W j Defined as representing the grid area G j The weight of the comprehensive importance is:

[0051] W j =af j +bC j +cL j

[0052] Among them, f j is the frequency of charging demand in the area, reflecting the concentration of low-power electric vehicles in the area;

[0053] C j It indicates the construction cost of the area, including land and construction costs;

[0054] L j is the load condition of the power grid in the area, considering the load pressure of the power grid in the area when it is connected to a new substation;

[0055] a, b, and c are important coefficients for adjusting the weights of each item.

[0056] Preferably, in step S3, formulating a dynamic weight adjustment strategy comprises the following steps:

[0057] S31: Periodically recalculate the demand frequency f based on the real-time driving data of the electric vehicle j ;

[0058] S32: Dynamically adjust the construction cost of the substation by dynamically evaluating the construction cost C j ;

[0059] S33: Power grid load status L j Conduct dynamic monitoring and adjustments.

[0060] Preferably, the step S3 further includes: iterating based on historical data, specifically:

[0061] As electric vehicle data and demand data continue to accumulate, the weights of each region are re-evaluated after each iteration. j ;

[0062] Reselect substations based on new data to ensure that substation layouts can cope with the latest changes in demand.

[0063] On the other hand, a selection system based on the above-mentioned active station area selection method based on the time-space transfer characteristics of electric vehicles is provided, characterized in that it includes:

[0064] The data collection and processing module is used to collect electric vehicle driving data and substation data, and pre-process the collected data to obtain the charging demand probability of the area;

[0065] The area selection and optimization module is used to: select and optimize the area for charging electric vehicles;

[0066] The dynamic weight adjustment strategy formulation module is used to: determine the weight in active station selection and formulate a dynamic weight adjustment strategy.

[0067] Compared with the prior art, the beneficial effects of the present invention are:

[0068] 1. A method to meet the dynamic charging needs of electric vehicles and actively select the research area is proposed. The operating characteristics of each vehicle in the large-scale electric vehicle dataset are considered to fully explore the spatiotemporal characteristics of large-scale electric vehicle charging and discharging;

[0069] 2. By using actual electric vehicle driving and charging and discharging data, we can select and optimize key substations to avoid local overloads, ensure the stability and reliability of the power grid, and implement electric vehicle charging and discharging scheduling strategies in the selected substations, which will help to effectively connect with renewable energy (such as solar energy), promote the full utilization of green energy, and support sustainable development goals;

[0070] 3. Based on real and effective electric vehicle driving and charging and discharging data, an alternative selection method is proposed, which uses the diversity exploration of K-center greed in graph density decay to solve the station selection problem. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 is a flow chart of the method of the present invention;

[0072] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0073] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the application equally.

[0074] In the present invention, terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention, and do not specifically refer to any part or element in the present invention and should not be understood as limitations on the present invention.

[0075] Example:

[0076] like Figure 1 As shown, this embodiment provides an active station area selection method based on the time-space transfer characteristics of electric vehicles, including the following steps:

[0077] S1: Collect electric vehicle driving data and substation data, and pre-process the collected data to obtain the charging demand probability of the area;

[0078] S2: Select and optimize the charging area for electric vehicles;

[0079] S3: Determine the weights in active station selection and develop a dynamic weight adjustment strategy.

[0080] In step S1, the collected electric vehicle driving data includes: timestamp, geographical location of the electric vehicle and SoC, and the driving data is stored as time series data D i，t ={t,x i ,y i , SoC i}, where x i represents the horizontal coordinate of the i-th car at time t, y i represents the vertical coordinate of the i-th car at time t, SoC i represents the SoC of the i-th car at time t;

[0081] The obtaining of the charging demand probability of the area includes:

[0082] S11: Divide all the stations into equal proportions, and each grid area G j The size is Δx×Δy, where Δx represents the horizontal coordinate of the area division size and Δy represents the vertical coordinate of the area division size.

[0083] S12: For each grid area G j, according to the number of electric vehicles entering the area and the proportion of low-power vehicles, calculate the charging demand probability P of the area j :

[0084]

[0085] Among them, N is the total number of electric vehicles in the area, M represents the total number of all divided areas, L j Represents the total number of charging times of the jth station;

[0086] Based on the charging demand probability P j The initial station selection strategy is to prioritize the area with the largest charging demand G j* As the first station:

[0087] j * = arg max j P j .

[0088] Step S2 comprises:

[0089] The possibility of charging an electric vehicle in a certain area is not only related to historical data, but also to the amount of electricity the electric vehicle consumes when driving in this area. j , its W j It represents the correction of the possibility of all electric vehicles charging in this area. According to the following formula, the greater the power of the electric vehicle, the lower the possibility of charging in this area. Taking into account the frequency of charging demand, the proportion of low power and the frequency of electric vehicles passing by, the charging possibility correction formula is:

[0090] W j =αP j +β(1-SoC i )

[0091] Among them, α and β are coefficients for adjusting each weight;

[0092] In order to calculate the charging possibility more accurately, several factors need to be considered:

[0093] First, historical charging data can provide us with a basic charging demand frequency, which is based on the statistics of past charging behaviors in the area.

[0094] Secondly, the low battery ratio of electric vehicles is a key factor, because vehicles are more likely to seek charging stations for charging when the battery is low;

[0095] Finally, the frequency with which electric vehicles pass through the station area is also important, because the more times a vehicle passes through, the corresponding charging demand will also increase;

[0096] Execute the weighted K-center greedy algorithm, including:

[0097] o i The value is (x i ,y i ), (x i ,y i ) represents the longitude and latitude of the i-th electric car, O represents the original data set, m represents the total number of points in the dataset; the graph method consists of the following steps:

[0098] 1. Construct a K nearest neighbor indicator matrix:

[0099] P = [p ij ] m×n

[0100] Among them, p ij Used to indicate o j Does it mean o i , if o j Indicates o i , then p ij =1, otherwise p ij =0;

[0101] Let p ij is replaced by the weights of the Gaussian kernel distance:

[0102]

[0103] W=[w ij ] m×n , w ij ∈[0, 1]

[0104] Among them, d x (x i ,y i ) refers to the Manhattan distance, and v represents all (x i ,y i ) is the variance of the distance between them;

[0105] 2. Normalize the weights of the nearest neighbor edges by a large number K to distinguish data points with small weight neighbors:

[0106] Gra(O)=[Gra(o 1 ),…,Gra(o j ),…,Gra(o m )] T ;

[0107] 3. After selecting the high-density data points, the neighboring points o j Try density decay:

[0108]

[0109] Gra(o j )=Gra(o j )-Gra(o i )·p ij , o j ∈neighbour(o i )

[0110] Among them, neighbor(o i ) gives the o obtained according to the matrix P i The neighbor point set, p i is the i-th row of P, Index(p i >0) gives the value that satisfies p i >0, based on the density of the graph, select the source with the highest density and iteratively reduce the density of neighboring sources;

[0111] The active selection of the substations can be described as follows: based on the driving history data of large-scale electric vehicles and the SoC of all current electric vehicles, several substations are found to represent all substations for simplified analysis;

[0112] To achieve this goal, it is necessary to select representative areas through optimization algorithms:

[0113] First, define a coverage radius, which means that any area within this radius can be regarded as the same representative area. Next, find a set of areas so that each area can cover as many other areas as possible while maximizing the coverage range.

[0114] To solve this problem, an iterative optimization method is used. In each iteration, we select a new area, calculate its coverage, and update the set of covered areas. The specific steps are as follows:

[0115] S221: Select an initial station area as a starting point and add it to the set of marked data points;

[0116] S222: In each iteration, the coverage of the current set of marked data points is calculated, and a new station area is selected to maximize the new coverage;

[0117] S223: Add the newly selected area to the set of marked data points and update the coverage:

[0118]

[0119] Among them, e 0 is the set of currently labeled data points, e 1is a set of selected labeled data points, δ describes e 1 For 0 The coverage radius is expressed in terms of e 0 A set of balls with a radius of δ as the center of each member can cover the entire e 1 , minimization is equivalent to the minimax station selection problem:

[0120]

[0121] Where Δ represents the calculated Euclidean distance, which is solved by using the 2-OPT solution, iterating the following equation:

[0122]

[0123] e∈e∪{o}

[0124] Where o is the data point selected in one iteration;

[0125] Steps S221 and S222 are repeatedly executed until all stations are calculated.

[0126] In the process of selecting the station area, the weight W j As a key factor, it is used to reflect the charging demand and priority of each area. However, charging demand is not the only factor to be considered. It must also be combined with external conditions such as grid load and substation construction cost. In order to more comprehensively evaluate the value of each substation, we need to introduce multiple influencing factors and assign appropriate weights to each factor. These factors may include the load level of the existing grid, future demand forecasts, construction costs, access capabilities of renewable energy, and policy orientation. The weight of each factor needs to be adjusted according to the actual situation to ensure that the final decision can meet various needs and constraints to the greatest extent. Therefore, step S3 includes:

[0127] W j Defined as representing the grid area G j The weight of the comprehensive importance is:

[0128] W j =af j +bC j +cL j

[0129] Among them, f j is the frequency of charging demand in the area, reflecting the concentration of low-power electric vehicles in the area;

[0130] C j It indicates the construction cost of the area, including land and construction costs;

[0131] Lj is the load condition of the power grid in the area, considering the load pressure of the power grid in the area when it is connected to a new substation;

[0132] a, b, and c are important coefficients for adjusting each weight;

[0133] In order to better adapt to the needs of different regions and changes in external conditions, the weights w need to be adjusted regularly. j Dynamic adjustment is a process that consists of the following steps:

[0134] 1. Regularly recalculate the demand frequency f j :

[0135] Demand frequency f j It reflects the charging demand in each area and can be updated based on the real-time driving data of electric vehicles;

[0136] 2. Dynamically adjust the construction cost of the substation area C j :

[0137] Construction cost j It is not a fixed value. It will change with the construction conditions, policy changes and market conditions of the substation. For example, rising land prices or increased material costs will affect the construction cost of the substation. By dynamically evaluating the construction cost, C is updated. j ;

[0138] 3. Dynamic monitoring and adjustment of power grid load j :

[0139] Grid load L j Directly affects the feasibility of access to the substation. If the grid load in a certain area is close to the upper limit, selecting this area as the substation may cause grid overload. Therefore, real-time monitoring of grid load is required to ensure that the selection of the substation does not bring unnecessary pressure to the grid. The load value L is updated regularly. j To reflect the real-time status of the power grid and dynamically adjust the selection of substations;

[0140] Iterative Optimization Strategy

[0141] In order to further improve the rationality of the area selection, it is necessary to introduce an iterative optimization strategy:

[0142] Iterate based on historical data:

[0143] As electric vehicle data and demand data continue to accumulate, the system re-evaluates the weights of each region after each iteration. j ;

[0144] Reselect substations based on new data to ensure that substation layouts can cope with the latest demand changes;

[0145] If the layout and weight of the area do not change much after several consecutive iterations, it means that the area layout has become stable;

[0146] By dynamically adjusting the weight W j ,The substation selection strategy can achieve flexible response to the ever-changing demand environment and ensure the rationality and sustainability of the substation layout.

[0147] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the described embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. An active station selection method based on the temporal and spatial transfer characteristics of electric vehicles, characterized in that: The following steps are involved: S1: Collect electric vehicle driving data and substation data, and pre-process the collected data to obtain the charging demand probability of the area; S2: Select and optimize the charging area for electric vehicles; S3: Determine the weights in active station selection and develop a dynamic weight adjustment strategy.

2. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 1 is characterized in that: In step S1, the collected electric vehicle driving data includes: timestamp, geographical location of the electric vehicle and SoC, and the driving data is stored as time series data D i，t ={t,x i ,y i , SoC i }, where x i represents the horizontal coordinate of the i-th car at time t, y i represents the vertical coordinate of the i-th car at time t, SoC i represents the SoC of the i-th car at time t; The obtaining of the charging demand probability of the area includes: S11: Divide all the stations into equal proportions, and each grid area G j The size is Δx×Δy, where Δx represents the horizontal coordinate of the area division size and Δy represents the vertical coordinate of the area division size. S12: For each grid area G j , according to the number of electric vehicles entering the area and the proportion of low-power vehicles, calculate the charging demand probability P of the area j : Among them, N is the total number of electric vehicles in the area, M represents the total number of all divided areas, L j Represents the total number of charging times for the jth station.

3. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 2 is characterized in that: The step S1 also includes: based on the charging demand probability P j The initial area selection strategy is to prioritize the area with the greatest charging demand. As the first station: I * max arg j P j .

4. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 2 is characterized in that: The step S2 comprises: S21: Correct the possibility of charging electric vehicles, specifically: for each area G j , its W j represents the correction of the possibility of all electric vehicles charging in this area, W j The calculation process is: W j =αP j +β(1-SoC i ) Among them, α and β are coefficients for adjusting each weight; S22: Execute the weighted K-center greedy algorithm, including: o i The value is (x i ,y i ), (x i ,y i ) represents the longitude and latitude of the i-th electric car, O represents the original data set, m represents the sum of the points in the data set; Construct a K nearest neighbor indicator matrix: Among them, p ij Used to indicate o j Does it mean o i , if o j Indicates o i , then p ij =1, otherwise p ij =0; Let p ij is replaced by the weights of the Gaussian kernel distance: In=[in ij ] m×n ,In ij ∈[0,1] Among them, d x (x i ,y i ) refers to the Manhattan distance, and v represents all (x i ,y i ) is the variance of the distance between them.

5. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 4 is characterized in that: Normalize the weights of the large number K of nearest neighbors: Game(O)=[Game(o1),…,Game(o j ),…,Game(about m )] T ; After selecting the high-density data points, the neighboring points o j Try density decay: neighbour(o i )={o e } e∈Ineex(pi＞0) Game(o j )=Game(o j )-Game(o i ) p ij ,about j ∈neighbor(o i ) Among them, neighbor(o i ) gives the o obtained according to the matrix P i The neighbor point set, p i is the i-th row of P, Index(p i >0) gives the value that satisfies p i > 0, based on the density of the graph, select the source with the highest density and iteratively reduce the density of neighboring sources.

6. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 5 is characterized in that: In step S22, iteratively reducing the density of neighbor sources includes: S221: Select an initial station area as a starting point and add it to the set of marked data points; S222: In each iteration, the coverage of the current set of marked data points is calculated, and a new station area is selected to maximize the new coverage; S223: Add the newly selected area to the set of marked data points and update the coverage: Among them, e 0 is the set of currently labeled data points, e 1 is a set of selected labeled data points, δ describes e 1 For 0 The coverage radius is expressed in terms of e 0 A set of balls with a radius of δ as the center of each member can cover the entire e 1 , minimization is equivalent to the minimax station selection problem: Where Δ represents the calculated Euclidean distance, which is solved by using the 2-OPT solution, iterating the following equation: e∈e∪{o} Where o is the data point selected in one iteration; Steps S221 and S222 are repeatedly executed until all stations are calculated.

7. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 6 is characterized in that: In step S3, the weight in active station area selection is determined as follows: W j Defined as representing the grid area G j The weight of the comprehensive importance is: W j =of j +bC j +cL j Among them, f j is the frequency of charging demand in the area, reflecting the concentration of low-power electric vehicles in the area; C j It indicates the construction cost of the area, including land and construction costs; L j is the load condition of the power grid in the area, considering the load pressure of the power grid in the area when it is connected to a new substation; a, b, and c are important coefficients for adjusting the weights of each item.

8. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 7 is characterized in that: In step S3, the formulation of a dynamic weight adjustment strategy includes the following steps: S31: Periodically recalculate the demand frequency f based on the real-time driving data of the electric vehicle j ; S32: Dynamically adjust the construction cost of the substation by dynamically evaluating the construction cost C j ; S33: Power grid load status L j Conduct dynamic monitoring and adjustments.

9. The active station area selection method based on the time-space transfer characteristics of electric vehicles according to claim 8 is characterized in that: The step S3 further includes: performing iteration based on historical data, specifically: As electric vehicle data and demand data continue to accumulate, the weights of each region are re-evaluated after each iteration. j ; Reselect substations based on new data to ensure that substation layouts can cope with the latest changes in demand.

10. A selection system based on the active station area selection method based on the time-space transfer characteristics of electric vehicles as claimed in claim 1, characterized in that: include: The data collection and processing module is used to collect electric vehicle driving data and substation data, and pre-process the collected data to obtain the charging demand probability of the area; The area selection and optimization module is used to: select and optimize the area for charging electric vehicles; The dynamic weight adjustment strategy formulation module is used to: determine the weight in active station selection and formulate a dynamic weight adjustment strategy.

Images