A method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles
By constructing a two-layer collaborative optimization framework and leveraging the schedulable characteristics of autonomous electric vehicles, the site selection and capacity determination of charging and discharging stations and the operation of the power distribution network are optimized. This solves the problem that existing technologies have failed to achieve collaborative optimization between charging and discharging stations and the power distribution network, thereby improving the system's economy and operational efficiency.
Patent Information
- Application Number
- CN202611182039.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-05
- Publication Date
- 2026-11-17
AI Technical Summary
Existing charging station planning methods fail to fully utilize the dispatchable characteristics of autonomous electric vehicles and fail to achieve two-level collaborative optimization of charging station site selection and capacity determination and distribution network operation, resulting in low overall system efficiency and exacerbating the peak-valley difference in distribution network load.
A two-layer collaborative optimization framework is constructed. By combining the upper-layer charging and discharging station planning model and the lower-layer charging and discharging load optimization scheduling model with power-transportation coupling constraints, the site selection, capacity determination and load allocation of charging and discharging stations are optimized. By utilizing the mobile energy storage characteristics of autonomous electric vehicles, the unified dynamic optimization of charging and discharging station layout and distribution network operation is achieved.
It significantly improves the economic efficiency of charging and discharging station planning schemes and the operating efficiency of power distribution networks, reduces network losses and smooths load fluctuations, and achieves two-way coordination between charging and discharging station layout and load allocation.
Smart Images

Figure CN123117598A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electric vehicle charging infrastructure planning and power distribution network operation optimization technology, and in particular to a method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles. Background Technology
[0002] Current charging station planning methods face numerous challenges. Traditional methods typically treat charging demand distribution as a fixed input that needs to be predicted using historical data, and only consider electric vehicles as unidirectional loads, ignoring their enormous potential as mobile energy storage participating in bidirectional interactions with the distribution network. Site selection and capacity determination are primarily aimed at minimizing investment and operating costs. Furthermore, traditional planning methods fail to fully utilize the characteristic that autonomous electric vehicle charging locations can be subject to unified scheduling, resulting in insufficient proactive control over the spatiotemporal distribution of charging loads. With the deepening of transportation and power grid coupling theory, multi-dimensional collaborative methods have gradually been introduced into the planning field. However, static configuration or unidirectional scheduling from a single perspective still limits the overall system efficiency and easily exacerbates local load peak-valley differences in the distribution network, threatening the safe operation of the power grid. To address the problem that traditional methods cannot coordinate the underlying power grid operation status, a two-layer collaborative approach combining charging station site selection and capacity determination with proactive spatial scheduling of autonomous electric vehicles has become a major research direction.
[0003] There have been several reports on the planning and scheduling of electric vehicle charging stations. For example, patent CN122114552A obtains a set of charging demand points by clustering electric vehicle trajectory data in the target area, and constructs a model to minimize the cost of charging facilities to optimize site selection and layout, with the constraint that all demand points are covered by at least one candidate station. Patent CN121552981A uses a preset weight evaluation model to determine vehicle charging priority and dynamically allocates real-time charging power within a rolling time window under the constraint of the total available power of the target charging station. Patent CN121638843A constructs a bipartite graph for the joint scheduling and charging of driverless taxi fleets, and uses a reinforcement learning model to dynamically obtain weights for matching decisions to maximize the long-term benefits of the platform. The aforementioned studies involve the optimization of charging station site selection costs, power allocation at a single site, and improvement of revenue on the traffic operation side. However, none of them have utilized the characteristic that autonomous electric vehicles can run empty in accordance with unified dispatch instructions to transform the static prediction of charging demand into active spatial allocation. They also failed to combine the time-sharing charging and discharging behavior of mobile energy storage with the nonlinear power flow optimization of the lower-level distribution network, and failed to achieve the two-level joint collaborative planning of upper-level site capacity configuration and lower-level distribution network multi-time period operation status.
[0004] Therefore, a site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles is needed. The upper layer aims to minimize the comprehensive annual cost, including network loss costs and the revenue from mobile energy storage participating in peak shaving and valley filling ancillary services, for charging and discharging station site selection and capacity determination. The lower layer, while maintaining the total charging and discharging demand in each time period, utilizes the schedulable characteristics of autonomous electric vehicles to spatially allocate charging and discharging loads. Coordination and optimization are achieved with the goal of minimizing the all-day network loss of the distribution network and the peak-valley difference of the regional load. Ultimately, this achieves unified dynamic optimization of charging and discharging station layout, bidirectional load allocation, and distribution network operation status. Summary of the Invention
[0005] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles. The present invention constructs a two-layer collaborative optimization framework to make unified decisions on the site selection and capacity determination of charging and discharging stations and the optimization of power distribution network operation, and fully leverages the regulation potential of bidirectional charging and discharging of mobile energy storage, thereby significantly improving the economy of the planning scheme and the operating efficiency of the power distribution network.
[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:
[0007] A method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles proposed according to the present invention includes:
[0008] S1. Construct a two-layer collaborative planning framework that considers the coupling of power and transportation, including an upper planning layer and a lower scheduling layer. The upper planning layer includes an upper charging and discharging station planning model, and the lower scheduling layer includes a lower charging and discharging load optimization scheduling model.
[0009] S2. The upper-level charging and discharging station planning model obtains the spatiotemporal distribution data of regional charging demand and urban road network land price information. Based on the spatiotemporal distribution data of regional charging demand, capacity constraints and demand accessibility constraints are constructed. Based on the spatiotemporal distribution data of regional charging demand and urban road network land price information, the comprehensive annual cost is initially calculated. Under the premise of meeting the capacity constraints, demand accessibility constraints and pile number range constraints, with the goal of minimizing the comprehensive annual cost, an optimization model is constructed that incorporates the revenue of mobile energy storage participating in grid ancillary services. The upper-level charging and discharging station planning model solves the optimization model to obtain the site, pile number and distribution network node information of the candidate charging and discharging station schemes, and then transmits them to the lower-level scheduling layer.
[0010] S3, the lower-level charging and discharging load optimization scheduling model uses the site location, number of piles, and distribution network node information of the candidate charging and discharging station schemes obtained from the upper-level charging and discharging station planning model as initial data. With the optimization objectives of minimizing the daily active power loss of the distribution network and minimizing the peak-valley difference of the regional load, it utilizes the characteristic of autonomous electric vehicles obeying unified scheduling instructions to establish power-traffic coupled spatial scheduling constraints including vehicle transfer time and driving energy consumption. Under the conditions of maintaining the conservation of total demand during the time period and satisfying the safety operation constraints of the distribution network and the power-traffic coupled spatial scheduling constraints, it realizes the spatial collaborative allocation of charging and discharging loads of mobile energy storage among charging and discharging stations, obtains the optimized daily active power loss of the distribution network, and feeds back the optimized daily active power loss of the distribution network to the upper-level planning layer.
[0011] S4. Iterate through S2 to S3 until the convergence condition is met. Record and output the site, number of piles, and distribution network node information of the candidate charging and discharging station with the best comprehensive annual cost, as the final site selection and capacity determination scheme.
[0012] As a further optimization scheme of the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, the planning model of the upper-level charging and discharging station in step S2 aims to minimize the comprehensive annual cost, specifically adopting the following objective function:
[0013] ;
[0014] in, Investment costs for the construction of charging and discharging stations; For the operation and maintenance costs of charging and discharging stations; Energy consumption cost of autonomous electric vehicles during empty runs; Opportunity loss costs for autonomous electric vehicles; The cost of waiting in line for autonomous electric vehicles; The cost of network loss reported back from the lower layer; Revenue from autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage; This is the total annual cost.
[0015] As a further optimization scheme of the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, the planning model of the upper-level charging and discharging station in step S2 satisfies the capacity constraint, the accessibility constraint of the demand point, the range constraint of the number of single-station piles, and the integer constraint of the access node.
[0016] Capacity constraint requirement The charging and discharging station is at the first Hourly comprehensive service intensity Less than 1:
[0017] ;
[0018] in, Indicates the first Each charging and discharging station, i.e., the number of the charging station; Indicates the first Hours, that is, 1 to 24 periods of a day;
[0019] This constraint also implicitly includes the number of piles configured per station. The lower limit requirement:
[0020] ;
[0021] in, , The first individual charging and discharging stations Number of times the autonomous electric vehicle reached charging and discharging per hour; , These represent the charging service rate and discharging service rate of a single charging pile, respectively.
[0022] Demand point accessibility constraints are calculated by determining the maximum distance that an autonomous electric vehicle can travel when forced charging is triggered. Defined, which satisfies:
[0023] ;
[0024] in, The state of charge threshold for triggering forced charging of autonomous electric vehicles; Battery capacity for autonomous electric vehicles; Energy consumption per unit distance;
[0025] For any point of demand, there must be at least one charging / discharging station that meets its spatial distance requirement:
[0026] ;
[0027] in, The spatial distance between the demand point and the charging / discharging station; For safety margin;
[0028] set up The minimum number of piles threshold, If the maximum number of piles is the threshold, then the range constraint for the number of piles is expressed as:
[0029] ;
[0030] The integer constraint on access nodes requires that the actual physical quantities involved in the planning of charging and discharging stations must be integers, including the number of charging piles at each charging and discharging station and the node number connected to the distribution network:
[0031] ;
[0032] in, It is a set of positive integers. For the first The node number of each charging and discharging station connected to the distribution network.
[0033] As a further optimization scheme for the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, the lower-level scheduling layer in step S3 takes the minimum active power loss of the distribution network throughout the day and the minimum peak-valley difference of the regional load as the optimization objectives, specifically adopting the following objective function:
[0034] ;
[0035] in, This refers to the active power loss of the network throughout the day; The peak-valley difference of the regional distribution network load; and This is the normalized baseline value; , These are the weighting coefficients; This is the value of the comprehensive objective function.
[0036] As a further optimization scheme for the site selection and capacity determination method of charging and discharging stations for autonomous electric vehicles described in this invention, the lower-level charging and discharging load optimization scheduling model in step S3 satisfies the node power balance constraint, line capacity constraint, total charging and discharging demand conservation constraint, voltage safety constraint, current over-limit constraint, node power constraint, power-transportation coupling constraint, and interactive power boundary constraint.
[0037] The node power balance constraint requires that the active and reactive power injected into each node satisfy Kirchhoff's laws:
[0038] ;
[0039] in, and They are nodes During the period The basic active and reactive loads; and They are nodes During the period The charging and discharging power of connected mobile energy storage; and These are the conductance and susceptance of the line where the branch circuit is located, respectively; and They are nodes With nodes The voltage amplitude; The voltage phase angle difference between nodes; For nodes With nodes A set of connected nodes;
[0040] The total charge and discharge demand conservation constraint requires that the lower-level space optimization strictly follow the total charge and discharge amounts given by the upper-level layer for each time period:
[0041] ;
[0042] in, Time period given to the upper layer Net total charging and discharging power in the region;
[0043] Line capacity constraints, voltage safety constraints, and current over-limit constraints are satisfied:
[0044] ;
[0045] ;
[0046] ;
[0047] in, and Branch roads in different time periods The active power transmitted and the current amplitude; , This is the upper limit of line power; and The lower and upper limits of safe voltage amplitude at distribution network nodes; The maximum allowable current value; For branch roads during time periods The reactive power transmission power;
[0048] Node power constraints are used to limit the range of active and reactive power injection at each node of the distribution network:
[0049] ;
[0050] in, and These are the active power and reactive power of the nodes, respectively. and These are the minimum and maximum limits for the active power of the nodes, respectively. and These are the minimum and maximum limits for the reactive power of the nodes, respectively.
[0051] The power-transportation coupling constraints include the time update equations and the energy state update equations for autonomous electric vehicles transitioning across nodes in the road network:
[0052] Assume that the autonomous electric vehicle starts from the original node Transferred to optimized nodes through unified scheduling Arrival time Reaching the state of charge Must meet:
[0053] ;
[0054] ;
[0055] in, and The autonomous electric vehicles left the original node. Time and state of charge; For the original node With optimization nodes The spatial distance between the road networks; Average vehicle speed; Energy consumption per unit distance; Battery capacity for autonomous electric vehicles; The state of charge threshold for triggering forced charging of autonomous electric vehicles;
[0056] The interactive power boundary constraint requires that the charging and discharging power of the bidirectional charging and discharging station be limited by the capacity of the distribution transformer to meet the following requirements:
[0057] ;
[0058] in, For the first The maximum grid-connected capacity of each charging and discharging station; and For respectively the first individual charging and discharging stations during time period The charging power and discharging power.
[0059] As a further optimization of the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, the calculation of queuing costs for autonomous electric vehicles considers the superposition of bidirectional charging and discharging tasks of mobile energy storage. The revenue of autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage adopts a comprehensive revenue model that includes time-of-use electricity price arbitrage and capacity compensation, specifically including:
[0060] The calculation of queuing costs for autonomous electric vehicles includes:
[0061] Calculate the charging service rate of a single charging pile separately. With discharge service rate It satisfies:
[0062] ;
[0063] in, , These are the charging power and discharging power of a single charging station, respectively. , These represent the average energy per charge and the average energy per discharge for autonomous electric vehicles, respectively.
[0064] Get the individual charging and discharging stations Number of times an autonomous electric vehicle reaches charging per hour and the number of times the discharge of autonomous electric vehicles arrives Calculate the overall service intensity including bidirectional services. :
[0065] ;
[0066] in, The number of charging piles at the charging and discharging station;
[0067] Will With charge and discharge cycles By incorporating the multi-service station exponential queuing M / M / c model, the calculations for the charging and discharging station are performed sequentially. hourly idle probability Average queue length and average waiting time :
[0068] ;
[0069] ;
[0070] ;
[0071] in, For summation index variables;
[0072] By combining the average waiting time and the number of charge / discharge cycles in each time period, the first... Total queuing time for each charging and discharging station throughout the day The total annual queuing cost is calculated by summing up all charging and discharging stations. :
[0073] ;
[0074] ;
[0075] in, This represents the total number of charging and discharging stations. This is the conversion factor for queuing costs per unit time.
[0076] The revenue calculation for autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage is as follows:
[0077] ;
[0078] in, and Time periods Discharge electricity price and charging electricity price; and Charging and discharging stations During the period Discharge power and charging power interacting with the power distribution network; and These represent the combined efficiency during the discharge and charging processes, respectively. For time intervals; The capacity compensation price for the ancillary services market; For charging and discharging stations Rated discharge capacity for participation in the capacity market; Revenue from autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage.
[0079] As a further optimization scheme for the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, in order to maintain the regional load peak-valley difference To leverage the convex optimization properties, auxiliary variables are introduced to linearize the nonlinear terms, specifically including:
[0080] Define peak-valley difference ,in and These represent the maximum and minimum values of the equivalent total active power load of the distribution network throughout the day;
[0081] Introduce auxiliary variables and apply the following linearization constraint inequality:
[0082] ;
[0083] in, For nodes During the period The basic active load; and They are nodes During the period The charging and discharging power of connected mobile energy storage.
[0084] As a further optimization scheme for the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, the lower-level charging and discharging load optimization scheduling model introduces a second-order cone relaxation method to transform the nonlinear power flow equation into a convex optimization problem during the solution process:
[0085] Introducing auxiliary variables , , as well as To replace the voltage magnitude and phase angle product term in the original nonlinear power flow equations, specifically satisfying:
[0086] ;
[0087] ;
[0088] ;
[0089] in, For nodes With nodes During the period The voltage phase angle difference; For nodes During the period The voltage amplitude; For nodes During the period The voltage amplitude;
[0090] To ensure the physical consistency between the newly introduced auxiliary variables and the voltage amplitude and phase angle, the following nonlinear equality constraint is established:
[0091] ;
[0092] Relax the nonlinear equality constraint into a second-order cone inequality constraint:
[0093] ;
[0094] After substituting the auxiliary variables into the lower-level charging and discharging load optimization scheduling model, the second-order cone programming solver is called to solve the problem, and the globally optimal spatial collaborative allocation of charging and discharging loads among charging and discharging stations is output.
[0095] As a further optimization scheme of the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, the upper-level charging and discharging station planning model uses an improved particle swarm optimization algorithm to solve the optimization model to obtain the site, number of piles and distribution network node information of the candidate charging and discharging station schemes.
[0096] As a further optimization scheme for the site selection and capacity determination method for charging and discharging stations of autonomous electric vehicles described in this invention, the upper-level charging and discharging station planning model in step S2 is solved using an improved particle swarm optimization algorithm, including:
[0097] In the improved particle swarm optimization algorithm, the position vector of each particle represents a candidate scheme for the location and capacity of a charging and discharging station. Each dimension of the position vector corresponds to a decision variable to be optimized, including the longitude coordinates, latitude coordinates, and number of charging piles of the charging and discharging station.
[0098] Employing a decreasing adaptive weight strategy, the first... Inertia weight of the next iteration Dynamically adjusts with the number of iterations:
[0099] ;
[0100] in, The initial maximum weight; The minimum weight at the end of the iteration; This represents the maximum number of iterations. This represents the current iteration number;
[0101] After updating particle position and velocity, a genetic operator is introduced to measure mutation probability. Applying Gaussian mutation to each dimension of the offspring particle, the mutated... Dimensional components for:
[0102] ;
[0103] in, These are random numbers distributed according to a standard normal distribution. For variable asynchronous length; and The first The upper and lower limits of the values of the decision variables are determined by the urban road network area boundary, the range of distribution network node numbers, and the upper limit of the number of piles that can be built at a single station. The coefficient of variation; For the offspring particle The values of the dimensional components before mutation. The total dimension of the particles is equal to the product of the number of charging stations and the number of variables to be optimized at each station.
[0104] Compared with the prior art, the present invention, employing the above technical solution, has the following technical effects:
[0105] (1) The present invention constructs a two-layer collaborative planning framework: the upper layer aims to minimize the comprehensive annual cost and decides the location and capacity of charging and discharging stations; the lower layer aims to minimize the distribution network loss and the peak-valley difference of regional load and optimizes the spatial allocation of charging and discharging loads. Through power-transport coupling constraints, it overcomes the defects of existing technologies that regard charging demand as a fixed input and ignore the bidirectional interactive capability and location schedulability of autonomous electric vehicles as mobile energy storage, and realizes the bidirectional collaboration between the layout of charging and discharging stations and the load allocation.
[0106] (2) The present invention adopts an improved particle swarm algorithm in the upper layer, which enhances the diversity of the population through adaptive inertial weights and genetic operators to avoid getting trapped in local optima; in the lower layer, auxiliary variable linearization and second-order cone relaxation are introduced to transform the non-convex nonlinear power flow problem containing peak-valley difference extreme values into a convex optimization problem, and the second-order cone programming solver is called to solve it. Under the premise of satisfying safety constraints, network loss is significantly reduced and load fluctuations are smoothed.
[0107] (3) This invention comprehensively considers the costs of construction investment, operation and maintenance, empty running energy consumption, opportunity loss, queuing, distribution network loss and mobile energy storage peak shaving and valley filling auxiliary service revenue, and coordinates the economic interests of the charging and discharging station investor and the distribution network operator. While meeting the charging and discharging needs of mobile energy storage, it achieves the optimal overall economic efficiency of the system. Attached Figure Description
[0108] Figure 1 A schematic diagram of the two-layer planning framework and interaction mechanism for a dedicated charging and discharging station considering vehicle scheduling provided by the present invention;
[0109] Figure 2 Flowchart of the improved particle swarm optimization algorithm for solving the upper-level charging and discharging station planning model provided by this invention;
[0110] Figure 3 Flowchart of the second-order cone programming solution for the lower-level charging load optimization scheduling model provided by this invention;
[0111] Figure 4 The above is a comparison chart of 24-hour network losses of the distribution network before and after optimization, provided by this invention.
[0112] Figure 5 The voltage distribution diagram of each node in the distribution network before optimization provided by this invention;
[0113] Figure 6 The optimized voltage distribution diagram of each node in the power distribution network provided by this invention. Detailed Implementation
[0114] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
[0115] This invention provides a method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles. First, it collects spatiotemporal distribution data of charging and discharging demand in the study area, urban road network land price information, and power distribution network parameters. Then, it constructs a two-layer planning framework consisting of an upper-layer charging and discharging station planning model and a lower-layer charging and discharging load optimization scheduling model. Figure 1As shown, the upper-level charging and discharging station planning model aims to minimize the overall annual cost, determining the geographical location of the charging and discharging stations and the number of bidirectional charging and discharging piles. The lower-level charging and discharging load optimization scheduling model aims to minimize the distribution network loss throughout the day and the peak-valley difference of the regional load, performing spatial optimization allocation of the charging and discharging load under the constraint of total charging and discharging power conservation. Finally, through the iterative collaboration between the upper and lower levels of the improved particle swarm optimization algorithm and the second-order cone programming solver, the optimal location and capacity sizing scheme is output.
[0116] In the two-tier planning framework, the upper planning layer is primarily responsible for the site selection and capacity determination decisions of charging and discharging stations, such as... Figure 2 As shown, this layer receives spatiotemporal distribution data of regional charging and discharging demand, and combines this data with constraints such as urban road network land price information and demand point accessibility to generate site, pile, and distribution network node information for candidate charging and discharging schemes. During the scheme evaluation process, the overall annual cost is considered, taking into account the investment in charging and discharging station construction, operation and maintenance, empty-run energy consumption, opportunity loss, queuing waiting time, and network loss costs fed back from lower layers. The revenue from mobile energy storage participating in peak shaving and valley filling ancillary services is also included. The site and capacity are optimized with the goal of minimizing the overall annual cost.
[0117] The lower-level scheduling layer, based on the location, capacity, and boundary conditions passed from the upper-level planning layer, is responsible for optimizing the spatial allocation of charging and discharging loads, such as... Figure 3 As shown, this invention fully utilizes the characteristic of autonomous electric vehicles that can obey unified scheduling commands. By scheduling autonomous electric vehicles to different charging and discharging stations, the originally uncertain allocation of charging and discharging demand is transformed into an optimizable decision-making process. Specifically, the lower-level scheduling layer, while keeping the total charging and discharging demand constant in each time period, spatially allocates the charging and discharging load borne by each station based on the location of the charging and discharging station access nodes, the station capacity limit, the distribution network topology parameters, basic load data, and network security operation constraints. Coordination and optimization are then performed with the goal of minimizing distribution network losses and minimizing the peak-valley difference in regional load.
[0118] In one embodiment, a method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles is provided, including the following control steps:
[0119] Step S1: This embodiment uses order data of autonomous electric ride-hailing vehicles in a certain city as a basis, dividing the study area into 180 regular hexagonal grids, and obtaining the predicted number of charging and discharging times and the net total charging and discharging power for each area at different time periods. Based on urban land price tier data, the study area is divided into three levels: Level 1 land price 38,850 yuan / ㎡, Level 2 land price 26,700 yuan / ㎡, and Level 3 land price 12,450 yuan / ㎡. The distribution network adopts the IEEE 33-node system, with a system baseline capacity of 100MVA. The charging and discharging loads of the above grids are mapped to the corresponding nodes of the distribution network, and the safe range of the distribution network voltage per unit value is set to 0.95 to 1.05 pu. The charging and discharging stations are set to use unmanned automatic bidirectional charging and discharging piles with robotic arms, with a unit price of 300,000 yuan / pile. Other simulation parameters are initialized as follows: discount rate 5%, depreciation period 20 years, maintenance cost ratio 1%, empty-run energy consumption cost 0.3 yuan / km, safety margin 1.1, single-station bidirectional charging and discharging pile range 1 to 100; single-vehicle battery capacity 60kWh, comprehensive charging and discharging efficiency 95%; peak shaving and valley filling time-of-use electricity price is set at 1.2, 0.7, and 0.3 yuan / kWh for peak, flat, and valley periods respectively, and ancillary service capacity compensation price is 300 yuan / kW·year; the upper-layer improved particle swarm algorithm is set with a population size of 50, a maximum number of iterations of 200, adaptive inertia weight reduced from 0.7 to 0.4, and Gaussian mutation probability of 0.8.
[0120] Step S2: Based on the acquired spatiotemporal distribution data of regional charging and discharging demand and urban road network land price information, construct an upper-level charging and discharging station planning model with the objective of minimizing the comprehensive annual cost. This model is used to generate site locations, number of charging piles, and distribution network node information for candidate charging and discharging schemes, and to pass the site selection and capacity determination schemes to the lower-level scheduling layer.
[0121] The objective of the upper-level charging and discharging station planning model is to minimize the overall annual cost. The objective function is as follows:
[0122]
[0123] in, Investment costs for the construction of charging and discharging stations; For the operation and maintenance costs of charging and discharging stations; Energy consumption cost of autonomous electric vehicles during empty runs; Opportunity loss costs for autonomous electric vehicles; The cost of waiting in line for autonomous electric vehicles; The cost of network loss reported back from the lower layer; Revenue from autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage; For the overall annual cost;
[0124] The specific calculation methods for the above costs and benefits are as follows:
[0125] (1) Investment cost of charging and discharging station construction This refers to the average annual cost of the one-time construction investment in a charging and discharging station, calculated using the annual value method. It includes land acquisition costs, charging pile purchase costs, and other supporting facility costs. The calculation formula is as follows:
[0126]
[0127] in, This represents the total number of planned charging and discharging stations. The discount rate; The depreciation period;
[0128] No. One-time investment and construction cost of a charging and discharging station Land acquisition fee Charging pile purchase cost Other supporting facility fees Cumulative composition:
[0129]
[0130]
[0131]
[0132] in, For the first The unit land cost of the area where a charging and discharging station is located; For the first The number of bidirectional charging and discharging piles configured in each charging and discharging station; The price of a single bidirectional charging / discharging pile; This is a quadratic conversion factor for other facility costs;
[0133] (2) Operation and maintenance costs of charging and discharging stations It is usually estimated as a certain percentage of the construction investment:
[0134]
[0135] in, This represents the proportion of operation and maintenance costs.
[0136] (3) Energy consumption cost of autonomous electric vehicles during empty runs Opportunity Loss Costs of Autonomous Electric Vehicles This reflects the loss of electrical energy, wear and tear, and time opportunities during the vehicle's empty journey to the charging and discharging station. The calculation formula is as follows:
[0137]
[0138]
[0139] in, This is the benchmark value for electricity prices; Energy consumption per unit distance; Cost of vehicle wear and tear per unit mileage; Total empty driving distance; The average revenue per unit time for the vehicle; Average vehicle speed;
[0140] (4) Queueing costs for autonomous electric vehicles Considering the superposition of bidirectional charging and discharging tasks for mobile energy storage, a time-segmented M / M / c queuing model is used for calculation:
[0141] First, calculate the charging service rate and discharging service rate of a single bidirectional charging and discharging pile:
[0142]
[0143] in, , These are the charging power and discharging power of a single charging station, respectively. , These represent the average energy per charge and the average energy per discharge for autonomous electric vehicles, respectively.
[0144] Get the individual charging and discharging stations Number of times an autonomous electric vehicle reaches charging per hour and the number of times the discharge of autonomous electric vehicles arrives Calculate the overall service intensity including bidirectional services. :
[0145]
[0146]
[0147] in, The number of charging piles at the charging and discharging station;
[0148] The service intensity With charge and discharge cycles Substitute the M / M / c queuing model and calculate the charging and discharging station sequentially. hourly idle probability Average queue length and average waiting time :
[0149]
[0150]
[0151]
[0152] in, For summation index variables;
[0153] By combining the average waiting time and the number of charge / discharge cycles in each time period, the first... Total queuing time for each charging and discharging station throughout the day The total annual queuing cost is calculated by summing up all charging and discharging stations. :
[0154]
[0155]
[0156] in, This represents the total number of charging and discharging stations. This is the discount factor for queuing costs per unit time.
[0157] (5) Network loss cost of lower layer feedback Daily power loss fed back from the lower-level distribution network optimization scheduling model Multiply by the electricity price to calculate the annual network loss cost:
[0158]
[0159] (6) Revenue from autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage A comprehensive revenue model incorporating time-of-use pricing arbitrage and capacity compensation is adopted:
[0160]
[0161] in, and Time periods Discharge electricity price and charging electricity price; and Charging and discharging stations During the period Discharge power and charging power interacting with the power distribution network; and These represent the combined efficiency during the discharge and charging processes, respectively. For time intervals; The capacity compensation price for the ancillary services market; For charging and discharging stations Rated discharge capacity for participation in the capacity market; Revenue from autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage;
[0162] The following constraints must be met when generating and updating particle planning schemes:
[0163] To ensure that charging and discharging stations can serve all arriving vehicles in a timely manner and avoid endless queues due to insufficient service capacity, the service intensity of each charging and discharging station must be less than 1 in any given hour.
[0164]
[0165] in, Indicates the first Each charging and discharging station, i.e., the number of the charging station; Indicates the first Hours, that is, 1 to 24 periods of a day;
[0166] According to the definition of service strength, capacity constraint is expressed as: This constraint also implicitly sets a lower limit on the number of piles that can be configured per station:
[0167]
[0168] in, , The first individual charging and discharging stations Number of times the autonomous electric vehicle reached charging and discharging per hour; , These represent the charging service rate and discharging service rate of a single charging pile, respectively.
[0169] To ensure that every charging need is effectively met, vehicles must have sufficient charge to reach the nearest charging station when forced charging is triggered. The maximum distance an autonomous electric vehicle can currently travel when forced charging is triggered. satisfy:
[0170]
[0171] in, The state of charge threshold for triggering forced charging of autonomous electric vehicles; Battery capacity for autonomous electric vehicles; Energy consumption per unit distance;
[0172] Introducing safety margin For any point of demand, there is at least one charging / discharging station. Satisfy its spatial distance Accessibility constraints:
[0173]
[0174] in, The spatial distance between the demand point and the charging / discharging station; For safety margin;
[0175] The number of charging piles at each charging and discharging station must be within a reasonable range of the project's and node's load-bearing capacity:
[0176]
[0177] in, The minimum number of piles threshold; The maximum number of piles threshold;
[0178] The actual physical quantities used in planning decisions must be integers:
[0179]
[0180] in, It is a set of positive integers. For the first The node number of each charging and discharging station connected to the power distribution network;
[0181] The aforementioned upper-level planning model is solved iteratively using an improved particle swarm optimization algorithm. Each particle is described by two vectors: velocity and position. In the nth iteration The particle in the first The speed and position update methods for the dimension are as follows:
[0182]
[0183]
[0184] in, For the first The inertia weight of the next iteration; and For learning factors; and A velocity perturbation used to increase the randomness of the optimization process; For the first The optimal position of each individual particle; This is the globally optimal position for the entire population;
[0185] The algorithm employs a decreasing adaptive weight strategy, which dynamically adjusts the inertia weight with the number of iterations:
[0186]
[0187] in, The initial maximum weight; The minimum weight at the end of the iteration; This represents the maximum number of iterations. This represents the current iteration number;
[0188] After updating velocity and position, the algorithm introduces a genetic operator to enhance population diversity. A roulette wheel selection method is used to choose two parent particles based on their fitness values. and Randomly select an intersection point The two parent particles are generated by performing a single-point intersection on the portion of the two parent particles after this point. and :
[0189]
[0190]
[0191] Perform Gaussian mutation on the offspring particles generated by the crossover, with the mutation probability Apply a Gaussian perturbation to each dimension of the variable, and the mutated 3rd... The dimensional components are:
[0192]
[0193] in, These are random numbers distributed according to a standard normal distribution. For variable asynchronous length; and The first The upper and lower limits of the values of the decision variables are determined by the urban road network area boundary, the range of distribution network node numbers, and the upper limit of the number of piles that can be built at a single station. The coefficient of variation; For the offspring particle The values of the dimensional components before mutation. The total dimension of the particles is equal to the product of the number of charging stations and the number of variables to be optimized at each station.
[0194] Step S3: Based on the given location and capacity boundary conditions of the charging and discharging stations in the upper layer, a lower-level scheduling model is constructed with the comprehensive objectives of minimizing the active power loss of the distribution network throughout the day and minimizing the peak-valley difference of the regional load. This model optimizes the spatial allocation of charging and discharging loads by guiding unmanned mobile energy storage vehicles to connect to different nodes while maintaining the conservation of total charging and discharging demand in each time period.
[0195] The objective of the lower-level scheduling model is to minimize the active power loss of the distribution network throughout the day and the peak-to-valley difference of the regional load. The objective function is constructed as follows:
[0196]
[0197] in, This refers to the active power loss of the network throughout the day; The peak-valley difference of the regional distribution network load; and This is the normalized baseline value; , These are the weighting coefficients; The value of the comprehensive objective function;
[0198] Active power loss of distribution network throughout the day The calculation formula is:
[0199]
[0200] in, For the distribution network in Active power loss during a given time period; For the set of all branches; For nodes With nodes Branch roads between; For branch circuit conductance; and They are nodes With nodes exist Voltage amplitude over a given period; The voltage phase angle difference between nodes; This is the system's baseline capacity;
[0201] Regional load peak and valley Defined as the maximum value of total active power load throughout the day. and minimum value The difference, that is To maintain the convexity of subsequent optimizations, auxiliary variables are introduced and the following linearization inequality constraint is applied to the extremum term:
[0202]
[0203] in, For nodes During the period The basic active load; and They are nodes During the period The charging and discharging power of connected mobile energy storage;
[0204] In the process of solving and optimizing the lower-level spatial collaborative scheduling model, the following physical operation constraints and power-transportation coupling constraints must be strictly satisfied:
[0205] (1) The active and reactive power injected into each node must satisfy Kirchhoff's laws:
[0206]
[0207] in, and They are nodes During the period The basic active and reactive loads; and They are nodes During the period The charging and discharging power of connected mobile energy storage; and These are the conductance and susceptance of the line where the branch circuit is located, respectively; and They are nodes With nodes The voltage amplitude; The voltage phase angle difference between nodes; For nodes With nodes A set of connected nodes;
[0208] (2) The line capacity constraint, voltage safety constraint, and current over-limit constraint are satisfied:
[0209]
[0210]
[0211]
[0212] in, and Branch roads in different time periods The active power transmitted and the current amplitude; , This is the upper limit of line power; and The lower and upper limits of safe voltage amplitude at distribution network nodes; The maximum allowable current value; For branch roads during time periods The reactive power transmission power;
[0213] (3) The total charging and discharging demand conservation constraint must be satisfied. The total charging and discharging power undertaken by the distribution network nodes in each time period must be equal to the net charging and discharging power value of the time period area given by the upper layer. :
[0214]
[0215] (4) Node power constraints are used to limit the range of active and reactive power injection at each node of the distribution network:
[0216]
[0217] in, and These are the active power and reactive power of the nodes, respectively. and These are the minimum and maximum limits for the active power of the nodes, respectively. and These are the minimum and maximum limits for the reactive power of the nodes, respectively.
[0218] (5) The power-traffic coupling constraints include the time update equation and the power state update equation for vehicle transfers across nodes in the road network. Assume that the autonomous electric vehicle starts from the original node... Transferred to optimized nodes through unified scheduling Arrival time Reaching the state of charge Must meet:
[0219]
[0220]
[0221] in, and The autonomous electric vehicles left the original node. Time and state of charge; For the original node With optimization nodes The spatial distance between the road networks; Average vehicle speed; Energy consumption per unit distance; Battery capacity for autonomous electric vehicles; The state of charge threshold for triggering forced charging of autonomous electric vehicles;
[0222] (6) The interactive power boundary constraint requires that the charging and discharging power of the bidirectional charging and discharging station be limited by the capacity of the distribution transformer and must meet the following requirements:
[0223] ;
[0224] in, For the first The maximum grid-connected capacity of each charging and discharging station; and For respectively the first individual charging and discharging stations during time period Charging power and discharging power;
[0225] Because the above model contains non-convex nonlinear power flow equations resulting from the product of voltage magnitude and phase angle, direct solution is difficult and it is hard to guarantee global optimality. A second-order cone relaxation method is introduced to transform the original non-convex problem into a convex optimization form, defining auxiliary variables... , , as well as To replace the product term in the original equation:
[0226]
[0227]
[0228]
[0229] in, , They are nodes With nodes During the period voltage amplitude, For nodes With nodes The voltage phase angle difference between them;
[0230] To ensure physical consistency, auxiliary variables must satisfy nonlinear equality constraints. To optimize the path through convexity, this equality constraint is further relaxed to a second-order cone inequality constraint:
[0231]
[0232] Substituting the auxiliary variables defined above into the lower-level objective function and Kirchhoff node power balance constraints, the original non-convex optimization problem is transformed into a standard second-order cone programming model. The second-order cone programming solver is then invoked to obtain a charging / discharging load spatial allocation scheme that minimizes the comprehensive evaluation objective. The corresponding active power losses for each time period are then fed back to the upper layer for calculating network loss costs to support the next round of collaborative iteration.
[0233] Step S4: Feed back the daily network loss power output from the lower layer to the upper layer to update the network loss cost item, and iteratively execute steps S2 to S3 for collaborative optimization until the algorithm converges. Perform the two-layer collaborative optimization on all schemes with 2 to 12 charging and discharging stations, record and compare the optimal comprehensive annual cost of each scheme, and finally select the scheme with the minimum comprehensive annual cost as the globally optimal charging and discharging station location, capacity setting and collaborative scheduling scheme for output.
[0234] Calculations show that when four charging and discharging stations are planned and constructed, the system's comprehensive annual cost reaches a minimum of 4.7452 million yuan. The cost details for different numbers of charging and discharging stations are shown in Table 1. The final optimal site selection and capacity allocation scheme for the charging and discharging stations includes the site coordinates, access to the distribution network nodes, and the number of bidirectional charging and discharging piles, as shown in Table 2.
[0235] Table 1. Detailed Cost Breakdown of the Number of Charging and Discharging Stations
[0236]
[0237] Table 2 Optimal Charging and Discharging Station Planning Scheme Configuration Table
[0238]
[0239] After optimization of the lower-level model, the total daily network loss of the distribution network decreased from 3591.09 kWh to 2918.37 kWh, a reduction of 18.73%; the minimum voltage at each node increased from 1.0292 pu to 1.0333 pu, the voltage range narrowed by 19.71%, and the voltage standard deviation decreased by 7.55%. The 24-hour network loss change curves before and after optimization are shown below. Figure 4 The voltage distribution at each node of the distribution network before and after optimization is shown in the figure. Figure 5 and Figure 6 .
[0240] In summary, this invention constructs a two-layer collaborative optimization framework, transforming the charging and discharging load from a fixed predictive input into a proactively controllable spatiotemporal decision variable, thus achieving unified optimization of charging and discharging station layout and distribution network operation. The lower layer introduces second-order cone relaxation technology, effectively reducing distribution network losses and improving voltage quality. The upper layer employs an improved particle swarm optimization algorithm, enhanced by adaptive inertial weights and genetic operators, avoiding local optima and obtaining an economically optimal planning scheme. This invention provides an effective solution for planning charging and discharging infrastructure for autonomous electric vehicles as a large-scale mobile energy storage bidirectional interactive scenario.
[0241] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.
Claims
1. A method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles, characterized in that, include: S1. Construct a two-layer collaborative planning framework that considers the coupling of power and transportation, including an upper planning layer and a lower scheduling layer. The upper planning layer includes an upper charging and discharging station planning model, and the lower scheduling layer includes a lower charging and discharging load optimization scheduling model. S2. The upper-level charging and discharging station planning model obtains the spatiotemporal distribution data of regional charging demand and urban road network land price information. Based on the spatiotemporal distribution data of regional charging demand, capacity constraints and demand accessibility constraints are constructed. Based on the spatiotemporal distribution data of regional charging demand and urban road network land price information, the comprehensive annual cost is initially calculated. Under the premise of meeting the capacity constraints, demand accessibility constraints and pile number range constraints, with the goal of minimizing the comprehensive annual cost, an optimization model is constructed that incorporates the revenue of mobile energy storage participating in grid ancillary services. The upper-level charging and discharging station planning model solves the optimization model to obtain the site, pile number and distribution network node information of the candidate charging and discharging station schemes, and then transmits them to the lower-level scheduling layer. S3, the lower-level charging and discharging load optimization scheduling model uses the site location, number of piles, and distribution network node information of the candidate charging and discharging station schemes obtained from the upper-level charging and discharging station planning model as initial data. With the optimization objectives of minimizing the daily active power loss of the distribution network and minimizing the peak-valley difference of the regional load, it utilizes the characteristic of autonomous electric vehicles obeying unified scheduling instructions to establish power-traffic coupled spatial scheduling constraints including vehicle transfer time and driving energy consumption. Under the conditions of maintaining the conservation of total demand during the time period and satisfying the safety operation constraints of the distribution network and the power-traffic coupled spatial scheduling constraints, it realizes the spatial collaborative allocation of charging and discharging loads of mobile energy storage among charging and discharging stations, obtains the optimized daily active power loss of the distribution network, and feeds back the optimized daily active power loss of the distribution network to the upper-level planning layer. S4. Iterate through S2 to S3 until the convergence condition is met. Record and output the site, number of piles, and distribution network node information of the candidate charging and discharging station with the best comprehensive annual cost, as the final site selection and capacity determination scheme.
2. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 1, characterized in that, In step S2, the planning model for the upper-level charging and discharging station aims to minimize the overall annual cost, specifically using the following objective function: ; in, Investment costs for the construction of charging and discharging stations; For the operation and maintenance costs of charging and discharging stations; Energy consumption cost of autonomous electric vehicles during empty runs; Opportunity loss costs for autonomous electric vehicles; The cost of waiting in line for autonomous electric vehicles; The cost of network loss reported back from the lower layer; Revenue from autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage; This is the total annual cost.
3. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 1, characterized in that, In step S2, the planning model for the upper-level charging and discharging station satisfies capacity constraints, accessibility constraints for demand points, range constraints for the number of single-station piles, and integer constraints for access nodes. Capacity constraint requirement The charging and discharging station is at the first Hourly comprehensive service intensity Less than 1: ; in, Indicates the first Each charging and discharging station, i.e., the number of the charging station; Indicates the first Hours, that is, 1 to 24 periods of a day; This constraint also implicitly includes the number of piles configured per station. The lower limit requirement: ; in, , The first individual charging and discharging stations Number of times the autonomous electric vehicle reached charging and discharging per hour; , These represent the charging service rate and discharging service rate of a single charging pile, respectively. Demand point accessibility constraints are calculated by determining the maximum distance that an autonomous electric vehicle can travel when forced charging is triggered. Defined, which satisfies: ; in, The state of charge threshold for triggering forced charging of autonomous electric vehicles; Battery capacity for autonomous electric vehicles; Energy consumption per unit distance; For any point of demand, there must be at least one charging / discharging station that meets its spatial distance requirement: ; in, The spatial distance between the demand point and the charging / discharging station; For safety margin; set up The minimum number of piles threshold, If the maximum number of piles is the threshold, then the range constraint for the number of piles is expressed as: ; The integer constraint on access nodes requires that the actual physical quantities involved in the planning of charging and discharging stations must be integers, including the number of charging piles at each charging and discharging station and the node number connected to the distribution network: ; in, It is a set of positive integers. For the first The node number of each charging and discharging station connected to the distribution network.
4. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 1, characterized in that, In step S3, the lower-level dispatching layer takes minimizing the active power loss of the distribution network throughout the day and minimizing the peak-valley difference of the regional load as the optimization objectives, and specifically adopts the following objective function: ; in, This refers to the active power loss of the network throughout the day; The peak-valley difference of the regional distribution network load; and This is the normalized baseline value; , These are the weighting coefficients; This is the value of the comprehensive objective function.
5. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 1, characterized in that, In step S3, the lower-level charging and discharging load optimization scheduling model satisfies the following constraints: node power balance, line capacity, total charging and discharging demand conservation, voltage safety, current over-limit, node power, power-transportation coupling, and interactive power boundary. The node power balance constraint requires that the active and reactive power injected into each node satisfy Kirchhoff's laws: ; in, and They are nodes During the period The basic active and reactive loads; and They are nodes During the period The charging and discharging power of connected mobile energy storage; and These are the conductance and susceptance of the line where the branch circuit is located, respectively; and They are nodes With nodes The voltage amplitude; The voltage phase angle difference between nodes; For nodes With nodes A set of connected nodes; The total charge and discharge demand conservation constraint requires that the lower-level space optimization strictly follow the total charge and discharge amounts given by the upper-level layer for each time period: ; in, Time period given to the upper layer Net total charging and discharging power in the region; Line capacity constraints, voltage safety constraints, and current over-limit constraints are satisfied: ; ; ; in, and Branch roads in different time periods The active power transmitted and the current amplitude; , This is the upper limit of line power; and The lower and upper limits of safe voltage amplitude at distribution network nodes; The maximum allowable current value; For branch roads during time periods The reactive power transmission power; Node power constraints are used to limit the range of active and reactive power injection at each node of the distribution network: ; in, and These are the active power and reactive power of the nodes, respectively. and These are the minimum and maximum limits for the active power of the nodes, respectively. and These are the minimum and maximum limits for the reactive power of the nodes, respectively. The power-transportation coupling constraints include the time update equations and the energy state update equations for autonomous electric vehicles transitioning across nodes in the road network: Assume that the autonomous electric vehicle starts from the original node Transferred to optimized nodes through unified scheduling Arrival time Reaching the state of charge Must meet: ; ; in, and The autonomous electric vehicles left the original node. Time and state of charge; For the original node With optimization nodes The spatial distance between the road networks; Average vehicle speed; Energy consumption per unit distance; Battery capacity for autonomous electric vehicles; The state of charge threshold for triggering forced charging of autonomous electric vehicles; The interactive power boundary constraint requires that the charging and discharging power of the bidirectional charging and discharging station be limited by the capacity of the distribution transformer to meet the following requirements: ; in, For the first The maximum grid-connected capacity of each charging and discharging station; and For respectively the first individual charging and discharging stations during time period The charging power and discharging power.
6. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 2, characterized in that, The calculation of queuing costs for autonomous electric vehicles takes into account the superposition of bidirectional charging and discharging tasks of mobile energy storage. The revenue of autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage adopts a comprehensive revenue model that includes time-of-use electricity price arbitrage and capacity compensation, specifically including: The calculation of queuing costs for autonomous electric vehicles includes: Calculate the charging service rate of a single charging pile separately. With discharge service rate It satisfies: ; in, , These are the charging power and discharging power of a single charging station, respectively. , These represent the average energy per charge and the average energy per discharge for autonomous electric vehicles, respectively. Get the individual charging and discharging stations Number of times an autonomous electric vehicle reaches charging per hour and the number of times the discharge of autonomous electric vehicles arrives Calculate the overall service intensity including bidirectional services. : ; in, The number of charging piles at the charging and discharging station; Will With charge and discharge cycles By incorporating the multi-service station exponential queuing M / M / c model, the calculations for the charging and discharging station are performed sequentially. hourly idle probability Average queue length and average waiting time : ; ; ; in, For summation index variables; By combining the average waiting time and the number of charge / discharge cycles in each time period, the first... Total queuing time for each charging and discharging station throughout the day The total annual queuing cost is calculated by summing up all charging and discharging stations. : ; ; in, This represents the total number of charging and discharging stations. This is the conversion factor for queuing costs per unit time. The revenue calculation for autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage is as follows: ; in, and Time periods Discharge electricity price and charging electricity price; and Charging and discharging stations During the period Discharge power and charging power interacting with the power distribution network; and These represent the combined efficiency during the discharge and charging processes, respectively. For time intervals; The capacity compensation price for the ancillary services market; For charging and discharging stations Rated discharge capacity for participation in the capacity market; Revenue from autonomous electric vehicles participating in peak shaving and valley filling ancillary services as mobile energy storage.
7. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 5, characterized in that, To maintain the regional load peak-valley difference To leverage the convex optimization properties, auxiliary variables are introduced to linearize the nonlinear terms, specifically including: Define peak-valley difference ,in and These represent the maximum and minimum values of the equivalent total active power load of the distribution network throughout the day; Introduce auxiliary variables and apply the following linearization constraint inequality: ; in, For nodes During the period The basic active load; and They are nodes During the period The charging and discharging power of connected mobile energy storage.
8. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 5, characterized in that, In solving the lower-level charging and discharging load optimization scheduling model, a second-order cone relaxation method is introduced to transform the nonlinear power flow equations into a convex optimization problem. Introducing auxiliary variables , , as well as To replace the voltage magnitude and phase angle product term in the original nonlinear power flow equations, specifically satisfying: ; ; ; in, For nodes With nodes During the period The voltage phase angle difference; For nodes During the period The voltage amplitude; For nodes During the period The voltage amplitude; To ensure the physical consistency between the newly introduced auxiliary variables and the voltage amplitude and phase angle, the following nonlinear equality constraint is established: ; Relax the nonlinear equality constraint into a second-order cone inequality constraint: ; After substituting the auxiliary variables into the lower-level charging and discharging load optimization scheduling model, the second-order cone programming solver is called to solve the problem, and the globally optimal spatial collaborative allocation of charging and discharging loads among charging and discharging stations is output.
9. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 1, characterized in that, The upper-level charging and discharging station planning model uses an improved particle swarm optimization algorithm to solve the candidate charging and discharging station schemes and obtain the station site, number of piles, and distribution network node information.
10. The method for site selection and capacity determination of charging and discharging stations for autonomous electric vehicles according to claim 5, characterized in that, In step S2, the planning model for the upper-level charging and discharging station is solved using an improved particle swarm optimization algorithm, including: In the improved particle swarm optimization algorithm, the position vector of each particle represents a candidate scheme for the location and capacity of a charging and discharging station. Each dimension of the position vector corresponds to a decision variable to be optimized, including the longitude coordinates, latitude coordinates, and number of charging piles of the charging and discharging station. Employing a decreasing adaptive weight strategy, the first... Inertia weight of the next iteration Dynamically adjusts with the number of iterations: ; in, The initial maximum weight; The minimum weight at the end of the iteration; This represents the maximum number of iterations. This represents the current iteration number; After updating particle position and velocity, a genetic operator is introduced to measure mutation probability. Applying Gaussian mutation to each dimension of the offspring particle, the mutated... Dimensional components for: ; in, These are random numbers distributed according to a standard normal distribution. For variable asynchronous length; and The first The upper and lower limits of the values of the decision variables are determined by the urban road network area boundary, the range of distribution network node numbers, and the upper limit of the number of piles that can be built at a single station. The coefficient of variation; For the offspring particle The values of the dimensional components before mutation. The total dimension of the particles is equal to the product of the number of charging stations and the number of variables to be optimized at each station.
Citation Information
Patent Citations
Ordered charging method, device and equipment for electric vehicle and storage medium
CN121552981A
Method and device for joint dispatching and charging of unmanned taxi team
CN121638843A
Electric vehicle charging station site selection planning method, system, equipment and medium
CN122114552A
