Electric swap station planning method considering electric vehicle user queuing factor

Abstract

The invention discloses a battery swap station planning method considering an electric vehicle user queuing factor, and the method comprises the steps: taking the site selection and capacity configuration of a battery swap station as decision variables, and taking the minimization of investment cost and user queuing length as optimization targets, and constructing a battery swap station planning model; firstly, the battery replacement demand of the electric vehicle is analyzed based on a trip chain; secondly, a Dijkstra algorithm is adopted to realize service area division of the battery swap station, and the dynamic operation process of the battery swap station is deeply studied; and finally, an immune genetic algorithm (IGA) is used for solving the constructed model, and the constructed model is used for solving the model. The method can effectively optimize the layout of the battery swap station, reduces the waiting time of a user, and has a high engineering application value.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy technology, in particular to the technology of electric vehicle energy supply facility planning, specifically to a battery swapping station planning method considering user queuing factors, covering electric vehicle battery swapping demand analysis, battery swapping station site selection and capacity configuration optimization, battery swapping dynamic operation modeling and intelligent optimization solution, aiming to reduce user queuing waiting time and realize more efficient battery swapping station planning scheme. BACKGROUND

[0002] At present, there has been some research on electric vehicle charging and battery swapping station planning. The basic idea is to analyze user charging and battery swapping demand, combine existing road network and power grid structure, determine the location and capacity of charging facilities under the condition of meeting the safe, reliable and economic operation of the power system, and meet user demand.

[0003] However, there are still many challenges and limitations in the planning process. In existing research, scholars at home and abroad focus on the operation mode of battery swapping stations, and some literature discusses the centralized charging and unified distribution mode. In this mode, fully charged batteries are transported from charging stations to battery swapping stations, and battery swapping stations are only responsible for battery replacement. Although the centralized charging mode can reduce the operational complexity of battery swapping stations, it may increase construction costs and weaken operational flexibility. On the other hand, the planning period is often long-term, and in the planning process of battery swapping stations, users are an important factor affecting planning in addition to economic considerations. In fact, if the interests of users are ignored, there may be a problem of long queues at some sites.

[0004] For example, the existing technical document "Optimal design of battery swapping-based electrified public bus transit systems" considers factors such as battery capacity, charging power, and battery quantity, and proposes an optimization model based on the economic efficiency of building battery swapping stations according to the characteristics of battery swapping stations. However, when planning battery swapping stations, only from the perspective of investment, it may cause some sites to have a queue, resulting in a decline in the overall efficiency of the planning scheme.

[0005] Therefore, developing a reasonable battery swapping station capacity configuration scheme can reduce the investment and operating costs of operators while taking into account the queuing experience of electric vehicle users, which has important theoretical value and engineering application significance for the scientific layout of battery swapping stations. SUMMARY

[0006] The application aims to solve the technical problem that the existing electric vehicle battery swap station planning only focuses on the optimization of a single index and ignores the long queue at some stations and the reduction of system comprehensive benefits caused by the user queuing experience.

[0007] To solve the above technical problems, the application adopts the technical scheme of: A battery swap station planning method considering electric vehicle user queuing factors, comprising the following steps: Step 1: obtaining the corresponding battery swap demand relationship between the battery swap station and the demand node, and obtaining the EV battery swap demand; Step 2: obtaining the EV queuing number; Step 3: based on the EV battery swap demand obtained in step 1 and the EV queuing number obtained in step 2, constructing a site selection and capacity determination model of the battery swap station, and obtaining a planning scheme from the model.

[0008] In step 1, the corresponding relationship between the battery swap station and the demand node is obtained, and the specific method is: Based on the trip chain, the EV battery swap demand is obtained, the planning area is divided into residential, working and commercial areas, and the electric vehicle battery swap demand of each area in different time periods is counted to obtain the demand node and its distribution; for each candidate battery swap station position, the shortest path distance from the candidate battery swap station to each node is obtained, and the battery swap station service area is divided according to the distance; The EV battery swap demand prediction model is used to obtain the battery swap demand of the battery swap station in different time periods .

[0009] In step 1, the trip chain is specifically: The trip behavior of electric vehicle users is abstracted as a sequential transfer path composed of three types of functional areas, i.e. residential area H, working area W and commercial area B, wherein the starting point and the ending point of the trip chain are both residential area H, and the user behavior mode is described by defining a plurality of typical trip chain combinations.

[0010] In step 1, the specific form of the EV battery swap demand prediction model is: The probability of the starting point and the ending point under a specific trip chain type, wherein the probability of the starting point adopts a uniform distribution assumption, and the probability of the ending point is given different weights according to whether the starting point and the ending point are in the same area; the probability of the starting point of the first trip of EV to the ending point is obtained by the following expression: (1) (2) (3) (4) In the formula, ​is the probability that the EV travel chain belongs to the zth category; Under the condition of the zth type, the starting point of the first journey is probability; It is of type z and the starting point of the first trip is The end point is probability; is the introduced intermediate variable; , and Form sets for H, W and B respectively; , and is the number of nodes contained in the three zones; Combined with the truncated normal distribution of the trip start time, a spatiotemporal joint probability distribution function is formed; assuming that the first trip start time of the EV is satisfy ,journey The starting time and itinerary -1 is related to the end time, according to the itinerary -1 end time, you can get the itinerary Truncated probability distribution function of the start time; the corresponding probability density function is: (5); Where, Expressed as - The end time of a trip; is the marginal probability density function of the trip start time; is the cumulative distribution function of the trip start time, indicating that the start time is less than or equal to probability; Considering that the travel time and location of EVs are independent of each other, the spatiotemporal probability distribution is as follows: (6); (7); Where, is the time conditional probability; is the spatial conditional probability; Indicates the type of chain; Indicates the The starting point of the trip; Indicates the The end point of the journey; The demand for battery replacement is based on the dynamic decay process of the battery state of charge (SOC). The initial SOC value of the EV is satisfy , when the EVs completed the journey The SOC value after -1 is: (8); Where, For the itinerary -1 SOC value at the start time; is the power consumption, unit is kWh / 100km; For the itinerary -1 driving distance; is the rated capacity of the battery; In position Department Number of EVs requiring charging at any given time It can be expressed by the following expression: (9); (10); (11); Where, For location Department Total EV at the moment; When it is 1, it means EVs need to be charged, otherwise it is 0; For the EVs at a time Battery state of charge; The minimum threshold for EV to be charged is 0.15 in the present invention; Indicates the The total number of trips in the trip chain to which the EV belongs; Indicate itinerary driving distance; is the total number of EVs; , and Represents the set of trip link type, location and time respectively; Graph theory is introduced to describe the road network topology. The adjacency matrix and its components are obtained as follows: (12); (13); Where, 、 is the distance between two nodes; inf is infinity, indicating that the node and No connection; For nodes and the distance between them; The division model of the battery swap station service area is shown as follows: ; (15); Where, Represents a slave node To the battery swap station The shortest distance is obtained by Dijkstra algorithm; Representation node Go to other battery swap stations The shortest path distance, ; is the set of all road network nodes; is the set of candidate battery swap station nodes; is a road segment in the road network; It is the collection of road sections passed through in the process; For road sections length; Battery swap stations Demand It can be obtained by the following formula: (16); Where, For nodes the needs of the place; For battery swap stations The service area node set.

[0011] In step 2, the number of EVs in the queue is obtained as follows: Based on the battery swap demand analysis in Step 1, a dynamic battery swap operation model for the battery swap station is constructed. This model categorizes battery states into three categories: charging, fully charged, and depleted. Based on the "first-come, first-served" principle, it combines real-time battery status with the number of arriving users to derive a dynamic calculation formula for queue length, thereby quantifying user waiting time. The number of EV queues is obtained from the dynamic operation model of battery swapping at the battery swap station .

[0012] In step 2, the characteristics of the battery swap dynamic operation model of the battery swap station include: 1) EV battery replacement is carried out in chronological order on a first-come, first-served basis; 2) The battery status in each time period is divided into: charging state, fully charged state, and exhausted state; 3) Batteries that are exhausted during this period will be charged in the next period, and batteries that are fully charged during this period will be put into use in the next period.

[0013] The dynamic calculation formula of the queuing queue length is: (17); In the formula, is the battery swap station In -1 period waiting for the number of EVs; is the battery swap station In The number of EVs added in the period; is the time period -1 start of the battery swap station The number of full-power available batteries, which is obtained by the following formula: (18); In the formula, and respectively, the number of battery swaps completed by the battery swap station In The number of full-power batteries in the period, which is calculated by the following expression: (19); (20); In the formula, is the battery swap station In The number of EVs added in the period -1; is the battery swap station In The number of battery swaps completed in the period -2; The number of various states in each time period is as follows: (21).

[0014] In step 3, the site selection and capacity model of the battery swap station is constructed, with the minimum investment cost and the minimum total length of the user queue as the dual objective functions, and the immune genetic algorithm is used for iterative optimization solution under the constraint conditions of safe operation of the power grid, until the planning scheme converges is output; The optimal site selection scheme is obtained from the site selection and capacity model of the battery swap station The battery configuration of each site .

[0015] In step 3, the site selection and capacity model of the battery swap station is as follows: Taking the site selection and capacity of the battery swap station as the decision variable, a planning model is constructed to minimize the investment cost and the length of the user queue: (22); (23); (twenty four); Where, is the weight; is the investment cost; The maximum possible investment cost; The length of the queue for users; is the maximum possible queue length; is the discount rate, which is 0.1 in the present invention; For the planning cycle; Representation node Whether to build a battery swap station; The number of batteries provided; and They are the land cost of the battery swap station and the cost of a single battery; and are respectively the node set and time period set of the battery swap station to be planned; For battery swap stations exist -1 queue length; 1) The node voltage and branch power flow constraints are shown as follows: (25); Where, 、 and are the upper and lower limits of the node voltage and the maximum value of the branch current respectively; For nodes In time The voltage amplitude; For branch The current amplitude; 2) The battery number constraint is as follows: (26); Where: is the maximum number of batteries; 3) The number of node access restrictions is as follows: (27); Where: and Build upper and lower limits on the total number of battery swap stations for all nodes respectively.

[0016] In step 3, the solution process of the immune genetic algorithm is as follows: Step 1: Parameter initialization; including EV battery swap requirements, road network and power grid data; Step 2: Generate an antibody population and memory library; its population information includes the location variables and capacity of the battery swap station; Step 3: fitness calculation; the population information is substituted into the objective function formula (22) to obtain the target value F; Step 4: calculation of antibody affinity and concentration; Step 5: judging whether the antibody concentration is 1; if yes, output the result; otherwise, go to step 6; Step 6: selection, crossover and mutation operations are performed; Step 7: on the basis of step 6, antibody individuals are selected by a probability method, and a new individual is generated by randomly changing the size of a certain gene value; return to step 3 to recalculate the fitness and perform the next iteration.

[0017] Compared with the prior art, the present application has the following technical effects: 1) The present application meets the economic requirements while taking into account the user queuing behavior factor, avoiding the problem of serious queuing at some sites due to one-sided pursuit of cost minimization, thereby improving the battery replacement service efficiency and reducing the user waiting time; 2) The present application combines the actual travel law of electric vehicles, the road network structure and the battery replacement demand distribution to achieve more accurate site selection and capacity configuration, and improve the matching degree of infrastructure and actual demand; 3) The present application reduces the waiting time of users during peak hours through queuing modeling and optimization, improves the response speed and reliability of the battery replacement service, and effectively enhances the user satisfaction; 4) The present application reduces battery redundancy and resource waste through optimization of capacity configuration and site layout, helping operators to control the cost of investment and operation. BRIEF DESCRIPTION OF DRAWINGS

[0018] The present application will be further described below in conjunction with the drawings and examples: Figure 1 is a schematic diagram of the framework structure; Figure 2 is a schematic diagram of the road network topology structure; Figure 3 is a flowchart of the solution of the battery replacement station planning model based on IGA; Figure 4 is the road network structure diagram used in the example; Figure 5 is a time and space distribution diagram of EV battery replacement demand; Figure 6 is a planning result and service area division diagram of method 2; Figure 7 is a planning result and service area division diagram of method 1. DETAILED DESCRIPTION

[0019] A battery replacement station planning method considering the queuing factor of electric vehicle users, comprising the following steps: Step 1: Analyze EV battery swapping demand based on travel chain theory and divide the planning area into living, working, and commercial areas. Monte Carlo simulation is used to calculate the EV battery swapping demand in each area at different times, and to obtain demand nodes and their distribution. Using the adjacency matrix in graph theory to represent the road network structure, the Dijkstra algorithm is used to calculate the shortest path distance from each candidate battery swap station to each node. Based on this, the battery swap station service area is divided and the corresponding relationship between battery swap stations and demand nodes is clarified. The EV battery swap demand prediction model is used to obtain the battery swap demand at the battery swap station in different time periods. ; Step 2: Based on the battery swap demand analysis in Step 1, a dynamic model of battery state transitions and user queuing behavior within the station is constructed. The battery states are divided into three categories: charging, fully charged, and depleted. Based on the "first come, first served" principle, combined with the real-time battery state and the number of arriving users, a dynamic calculation formula for queue length is derived, thereby establishing a quantitative evaluation mechanism for user waiting time. The number of EV queues is obtained from the dynamic operation model of battery swapping at the battery swap station ; Step 3: Based on the EV battery swapping demand and the number of EV queues obtained in Steps 1 and 2, an optimization model is constructed with the location of the battery swap station and the number of battery configurations as decision variables. The dual objective function is to minimize the investment cost and the total length of the user queue. Under the constraints of satisfying the safe operation of the power grid, an immune genetic algorithm is used to perform iterative optimization and solve the problem until a convergent planning solution is output.

[0020] The optimal site selection scheme is obtained from the site selection and capacity determination model of the battery swap station Battery configuration at each site .

[0021] The travel chain is: The travel behavior of electric vehicle users is abstracted into a sequential transfer path consisting of three functional areas: living area (H), work area (W), and business area (B). The starting and ending points of the travel chain are both in the living area (H). Twelve typical travel chain combinations (such as simple chain HWH and complex chain HWBH) are defined to describe user behavior patterns. The 12 typical travel chain combinations are: Simple chain: HWH, HHH, HBH; Complex chain: HWBH, HBWH, HHWH, HWHH, HWWH, HHHH, HBHH, HBBH, HHBH; See Table 1 for details: Table 1 Travel chain combination table

[0022] The specific requirements for EV battery replacement are: Based on the trip chain theory, a probability model is constructed, and the joint probability distribution model is used to calculate the probability of the starting and ending points under a specific trip chain type. The starting point probability adopts the uniform distribution assumption, and the end point probability is given differentiated weights based on whether the starting and ending points are in the same area. To the end The probability of is given by the following expression: (1); (2); (3); (4); Where, is the probability that the EV travel chain belongs to the zth category; Under the condition of the zth type, the starting point of the first journey is probability; It is of type z and the starting point of the first trip is The end point is probability; is the introduced intermediate variable; , and Form sets for H, W and B respectively; , and is the number of nodes contained in the three zones.

[0023] Combined with the truncated normal distribution of the trip start time, a spatiotemporal joint probability distribution function is formed; assuming that the first trip start time of the EV is satisfy ,journey The starting time and itinerary -1 is related to the end time, according to the itinerary -1 end time, you can get the itinerary Truncated probability distribution function of the start time; the corresponding probability density function is: (5); Where, Expressed as - The end time of a trip; is the marginal probability density function of the trip start time; is the cumulative distribution function of the trip start time, indicating that the start time is less than or equal to probability.

[0024] Considering that the travel time and location of EVs are independent of each other, the spatiotemporal probability distribution is as follows: (6); (7); The basis for battery replacement demand is the dynamic decay process of the battery state of charge (SOC). The initial SOC value of the EV is satisfy , when the EVs completed the journey The SOC value after -1 is: (8); Where, For the itinerary -1 SOC value at the start time; is the power consumption, unit is kWh / 100km; For the itinerary -1 driving distance; is the rated capacity of the battery; In position Department Number of EVs requiring charging at any given time It can be expressed by the following expression: (9); (10); (11); Where, For location Department Total EV at the moment; When it is 1, it means EVs need to be charged, otherwise it is 0; For the EVs at a time Battery state of charge; The minimum threshold for EV to be charged is 0.15 in the present invention; Indicates the The total number of trips in the trip chain to which the EV belongs; Indicate itinerary driving distance; is the total number of EVs; , and Represents the set of trip link type, location and time respectively.

[0025] Graph theory is introduced to describe the road network topology. The adjacency matrix and its components are obtained as follows: (12); (13); Where, 、 is the distance between two nodes; inf is infinity, indicating that the node and No connection; For nodes and The distance between them.

[0026] The division model of the service area of ​​the battery swap station is shown as follows: (14); (15); Where, Represents a slave node To the battery swap station The shortest distance is obtained by Dijkstra algorithm; Representation node Go to other battery swap stations The shortest path distance, ; is the set of all road network nodes; is the set of candidate battery swap station nodes; is a road segment in the road network; It is the collection of road sections passed through in the process; For road sections length; Battery swap stations Demand It can be obtained by the following formula: (16); Where, For nodes the needs of the place; For battery swap stations The service area node set.

[0027] The dynamic operation model of battery swapping at a battery swap station follows the following principles: 1) EV battery replacement is carried out in chronological order on a first-come, first-served basis; 2) The battery status in each time period is divided into: charging state, fully charged state, and exhausted state.

[0028] 3) Batteries that are exhausted during this period will be charged in the next period, and batteries that are fully charged during this period will be put into use in the next period.

[0029] The formula for calculating the length of the EV queue is: (17); Where, For battery swap stations exist -1 period of waiting EV number; For battery swap stations exist The number of EVs added during the period; For time period -1 Switch station at the start The number of fully charged batteries available is obtained by the following formula: (18); Where, and Battery swap stations exist The number of battery swaps completed and the number of fully charged batteries during the time period are calculated using the following expressions: (19); (20); The constraints on the number of states in each time period are as follows: (twenty one); The site selection and capacity determination model of the battery swap station is specifically as follows: Taking the location and capacity of battery swap stations as decision variables, a planning model is constructed with the optimization objectives of minimizing investment costs and user queue lengths.

[0030] (twenty two); (twenty three); (twenty four); Where, is the weight; is the investment cost; The maximum possible investment cost; The length of the queue for users; is the maximum possible queue length; is the discount rate, which is 0.1 in the present invention; For the planning cycle; Representation node Whether to build a battery swap station; The number of batteries provided; and They are the land cost of the battery swap station and the cost of a single battery; and are respectively the node set and time period set of the battery swap station to be planned; For battery swap stations exist -1 queue length.

[0031] 1) The node voltage and branch power flow constraints are shown as follows: (25); Where, 、 and are the upper and lower limits of the node voltage and the maximum value of the branch current respectively; For nodes In time The voltage amplitude; For branch The current amplitude.

[0032] 2) The battery number constraint is as follows: (26); Where: The maximum number of batteries.

[0033] 3) The number of node access restrictions is as follows: (27); Where: and Build upper and lower limits on the total number of battery swap stations for all nodes respectively.

[0034] The solution process of the immune genetic algorithm is as follows: IGA is characterized by strong robustness and fast convergence, and it also combines biological immune thinking to achieve effectiveness in the iterative process. The specific steps are: Step 1: Parameter initialization, including EV battery swap requirements, road network and power grid data; Step 2: Generate an antibody population and memory library. Its population information includes the location variables and capacity of the battery swap station; Step 3: Fitness calculation. Substitute the population information into the objective function (22) to obtain the target value F; Step 4: Calculation of antibody affinity and concentration; Step 5: Determine whether the antibody concentration is 1. If so, output the result; otherwise, go to step 6; Step 6: Perform selection, crossover, and mutation operations; Step 7: Based on step 6, select antibody individuals using a probabilistic method and randomly change the size of a gene value to generate new individuals. Return to step 3 to recalculate fitness and proceed to the next round of iteration.

[0035] Example: The road network topology used in the example includes 51 roads and 33 road network nodes in the target area. The prices of functional zones B, W, and H are 120,000, 80,000, and 60,000 yuan / m2, respectively. The road lengths are shown in Table 2. The grid structure adopts IEEE 33 nodes, the planning period is 5 years, and the battery unit price is 100,000 yuan / unit.

[0036] Table 2 Road length information parameters

[0037] 1) Process simulation The EV battery replacement demand distribution and planning results are obtained through simulation as follows: Figure 4 and Figure 5 Detailed data on the spatiotemporal distribution of EV battery replacement demand is shown in Table 3.

[0038] Table 3 Spatiotemporal distribution data of EV battery replacement demand

[0039] Depend on Figure 5 It can be seen that battery swap stations are established at network nodes 6, 17, 29, 5, and 20. Among them, the battery swap station at node 6 is equipped with 28 batteries, and its service range is nodes 1, 2, 6, 7, 8, 11, 16, and 31; the battery swap station at node 17 is equipped with 18 batteries, and its service range is nodes 12, 17, 18, 21, and 22; the battery swap station at node 29 is equipped with 26 batteries, and its service range is nodes 25, 26, 27, 29, 30, and 33; the battery swap station at node 5 is equipped with 30 batteries, and its service range is nodes 3, 4, 5, 9, 10, and 32; the battery swap station at node 20 is equipped with 29 batteries, and its service range is nodes 13, 14, 15, 19, 20, 23, 24, and 28.

[0040] 2) Comparative Analysis In order to verify the necessity of the method proposed in the present invention, the following method is designed: Method 1: A BSS planning method that takes economy as the goal and does not consider user queuing; Method 2: A BSS planning method that comprehensively considers economy and user queuing, namely the method of the present invention.

[0041] The optimal planning solution obtained by method 1 is as follows Figure 6 As shown in the figure, the planning schemes obtained by the two methods and the normalization and weighting of the investment and queue length of methods 1 and 2 based on formula (1) (where =0.5, Take 500, Take 20) to obtain the target values ​​of the two methods As shown in Table 4: Table 4 Planning costs under two methods

[0042] As shown in Table 4, the investment cost of Method 2 increased by 131,900 yuan. The main reason is that Method 2 increased the number of battery swap stations, which directly increased its investment cost. From the perspective of queues, there was no queue at any station in Method 2. The reason is that compared with Method 1, Method 2 increased the number of battery swap stations so that users' battery swap requests were met in a timely manner, thus effectively reducing the length of the queue. From the perspective of the normalized index F, the target value F of Method 2 is smaller than that of Method 1. The reason is that compared with Method 1, although Method 2 has a higher investment cost, it did not have a queue. Therefore, Method 2 has a better advantage in comprehensive comparison.

Claims

1. A battery swap station planning method considering the queuing factors of electric vehicle users, characterized in that: The following steps are involved: Step 1: Obtain the corresponding battery swap demand relationship between the battery swap station and the demand node, and obtain the EV battery swap demand; Step 2: Get the number of EVs in queue; Step 3: Based on the EV battery swapping demand obtained in step 1 and the number of EV queues obtained in step 2, a site selection and sizing model for battery swap stations is constructed, and a planning scheme is obtained from the model.

2. The method according to claim 1, characterized in that In step 1, the corresponding relationship between the battery swap station and the demand node is obtained as follows: Based on the travel chain, the EV battery swapping demand is obtained, and the planning area is divided into living, working, and commercial areas. The EV battery swapping demand in each area at different times is counted to obtain the demand nodes and their distribution. For each candidate battery swap station location, the shortest path distance from the candidate battery swap station to each node is calculated, and the battery swap station service area is divided accordingly. The EV battery swap demand prediction model is used to obtain the battery swap demand at the battery swap station in different time periods. .

3. The method according to claim 1, characterized in that In step 1, the travel chain is specifically as follows: The travel behavior of electric vehicle users is abstracted into a sequential transfer path consisting of three functional areas: living area H, work area W, and commercial area B. The starting point and end point of the travel chain are both the living area H. The user behavior pattern is described by defining multiple typical travel chain combinations.

4. The method according to any one of claims 1 to 3, characterized in that In step 1, the specific form of the EV battery replacement demand prediction model is: The probability of the starting and ending points under a specific travel chain type, where the starting point probability adopts a uniform distribution assumption, and the end point probability is given differentiated weights based on whether the starting and ending points are in the same district; the starting point of the first EV trip To the end The probability of is given by the following expression: （1）； （2）； （3）； （4）； Where, is the probability that the EV travel chain belongs to the zth category; Under the condition of the zth type, the starting point of the first journey is probability; It is of type z and the starting point of the first trip is The end point is probability; is the introduced intermediate variable; , and Form sets for H, W and B respectively; , and is the number of nodes contained in the three zones; Combined with the truncated normal distribution of the trip start time, a spatiotemporal joint probability distribution function is formed; assuming that the first trip start time of the EV is satisfy ,journey The starting time and itinerary -1 is related to the end time, according to the itinerary -1 end time, you can get the itinerary Truncated probability distribution function of the start time; the corresponding probability density function is: （5）； Where, Expressed as - The end time of a trip; is the marginal probability density function of the trip start time; is the cumulative distribution function of the trip start time, indicating that the start time is less than or equal to probability; Considering that the travel time and location of EVs are independent of each other, the spatiotemporal probability distribution is as follows: （6）； （7）； Where, is the time conditional probability; is the spatial conditional probability; Indicates the type of chain; Indicates the The starting point of the trip; Indicates the The end point of the journey; The demand for battery replacement is based on the dynamic decay process of the battery state of charge (SOC). The initial SOC value of the EV is satisfy , when the EVs completed the journey The SOC value after -1 is: （8）； Where, For the itinerary -1 SOC value at the start time; is the power consumption, unit is kWh / 100km; For the itinerary -1 driving distance; is the rated capacity of the battery; In position Department Number of EVs requiring charging at any given time It can be expressed by the following expression: （9）； （10）； （11）； Where, For location Department Total EV at the moment; When it is 1, it means EVs need to be charged, otherwise it is 0; For the EVs at a time Battery state of charge; The minimum threshold for charging an EV; Indicates the The total number of trips in the trip chain to which the EV belongs; Indicate itinerary driving distance; is the total number of EVs; , and Represents the set of trip link type, location and time respectively; Graph theory is introduced to describe the topological structure of the road network. The adjacency matrix and its components are obtained as follows: （12）； （13）； Where, 、 is the distance between two nodes; inf is infinity, indicating that the node and No connection; For nodes and the distance between them; The division model of the battery swap station service area is shown as follows: （14）； （15）； Where, Represents a slave node To the battery swap station The shortest distance is obtained by Dijkstra algorithm; Representation node Go to other battery swap stations The shortest path distance, ; is the set of all road network nodes; is the set of candidate battery swap station nodes; is a road segment in the road network; It is the collection of road sections passed through in the process; For road sections length; Battery swap stations Demand It can be obtained by the following formula: （16）； Where, For nodes the needs of the place; For battery swap stations The service area node set.

5. The method according to claim 1, wherein In step 2, the number of EVs in the queue is obtained as follows: Based on the battery swap demand analysis in Step 1, a dynamic battery swap operation model for the battery swap station is constructed. This model categorizes battery states into three categories: charging, fully charged, and depleted. Based on the "first-come, first-served" principle, it combines real-time battery status with the number of arriving users to derive a dynamic calculation formula for queue length, thereby quantifying user waiting times. The number of EV queues is obtained from the dynamic operation model of battery swapping at the battery swap station .

6. The method according to claim 5, characterized in that In step 2, the characteristics of the battery swap dynamic operation model of the battery swap station include: 1) EV battery replacement is carried out in chronological order on a first-come, first-served basis; 2) The battery status in each time period is divided into: charging state, fully charged state, and exhausted state; 3) Batteries that are exhausted during this period will be charged in the next period, and batteries that are fully charged during this period will be put into use in the next period.

7. The method according to claim 5, characterized in that The dynamic calculation formula for the queue length is: （17）； Where, For battery swap stations exist -1 period of waiting EV number; For battery swap stations exist The number of EVs added during the period; For time period -1 Switch station at the start The number of fully charged batteries available is obtained by the following formula: （18）； Where, and Battery swap stations exist The number of battery swaps completed and the number of fully charged batteries during the time period are calculated using the following expressions: （19）； （20）； Where, For battery swap stations exist -The number of EVs added during period 1; For battery swap stations exist -2 Number of battery swaps completed within the time period; The constraints on the number of states in each time period are as follows: （21）。 8. The method according to claim 1, characterized in that In step 3, a site selection and sizing model for battery swap stations is constructed, with the dual objective function of minimizing investment cost and total length of user queues. Under the constraints of safe grid operation, an immune genetic algorithm is used to perform iterative optimization until a convergent planning solution is output. The optimal site selection scheme is obtained from the site selection and capacity determination model of the battery swap station Battery configuration at each site .

9. The method according to claim 8, characterized in that In step 3, the site selection and capacity determination model of the battery swap station is specifically as follows: Taking the location and capacity of the battery swap station as decision variables, a planning model is constructed with the optimization objectives of minimizing investment costs and user queue lengths: （22）； （23）； （24）； Where, is the weight; is the investment cost; The maximum possible investment cost; The length of the queue for users; is the maximum possible queue length; is the discount rate; For the planning cycle; Representation node Whether to build a battery swap station; The number of batteries provided; and They are the land cost of the battery swap station and the cost of a single battery; and are respectively the node set and time period set of the battery swap station to be planned; For battery swap stations exist -1 queue length; 1) The node voltage and branch power flow constraints are shown as follows: （25）； Where, 、 and are the upper and lower limits of the node voltage and the maximum value of the branch current respectively; For nodes In time The voltage amplitude; For branch The current amplitude; 2) The battery number constraint is as follows: （26）； Where: is the maximum number of batteries; 3) The number of node access restrictions is as follows: （27）； Where: and Build upper and lower limits on the total number of battery swap stations for all nodes respectively.

10. The method according to claim 9, characterized in that In step 3, the solution process of the immune genetic algorithm is as follows: Step 1: Parameter initialization; including EV battery swap requirements, road network and power grid data; Step 2: Generate an antibody population and memory library; its population information includes the location variables and capacity of the battery swap station; Step 3: Fitness calculation; Substitute the population information into the objective function (22) to obtain the target value F; Step 4: Calculation of antibody affinity and concentration; Step 5: Determine whether the antibody concentration is 1; if so, output the result; Otherwise go to step 6; Step 6: Perform selection, crossover, and mutation operations; Step 7: Based on step 6, select antibody individuals using a probabilistic method and randomly change the size of a gene value to generate new individuals; Return to step 3 to recalculate the fitness and proceed to the next round of iteration.

Images