Charging station and power distribution network multi-stage collaborative planning method oriented to electric vehicle load increase

Abstract

The invention belongs to the technical field of charging station planning, and relates to a charging station and power distribution network multi-stage collaborative planning method for electric vehicle load increase. The method comprises the following steps: initializing data; predicting the multi-stage inventory and the charging load of the electric vehicle; optimizing the site selection of the charging station; dynamic traffic flow path distribution; checking coupling operation of the traffic network and the power distribution network; and a charging station multi-stage planning decision is made. According to the invention, the problem that static single-stage planning cannot adapt to long-term load increase is solved; the problem of insufficient driving mileage caused by separation of station distribution and actual endurance demand of the electric vehicle is solved; the defect that a traditional static model cannot respond to real-time change of a road network is overcome; the technical effect that the total construction and operation cost is optimal is achieved; the station layout is highly matched with the driving distance, the cruising ability and the future space-time demand of the electric vehicle; a path with the optimal cost and charging node selection are provided for a user, and the overall operation efficiency and stability of a traffic and power grid coupling system are improved.

Description

Technical Field

[0001] This invention belongs to the field of charging station planning technology, and more specifically, it relates to a multi-stage collaborative planning method for charging stations and power distribution networks in response to the growth of electric vehicle load. Background Technology

[0002] Electric vehicles, as the core carrier of low-carbon travel, are being adopted at an increasingly rapid pace. However, there are many prominent problems in the current planning and operation of electric vehicle charging stations: (1) Uneven geographical distribution: The charging stations are overly concentrated in the core urban areas and sparsely distributed in the suburbs and remote areas, which leads to range anxiety when electric vehicles travel across regions, limiting the scope of use and promotion; (2) Insufficient number of charging piles: The rapid growth of electric vehicle ownership is not matched with the construction speed of charging piles. The supply and demand contradiction is prominent in commercial centers, residential areas and other areas, and the phenomenon of charging queues occurs frequently. (3) Significant cost constraints: The construction of charging stations involves high costs such as equipment procurement, site leasing and power access, and during operation, continuous expenses such as equipment maintenance, power consumption and personnel management must also be borne; (4) Insufficient planning adaptability: Existing plans are mostly based on static charging load distribution, which cannot adapt to the long-term dynamic growth of the number of electric vehicles and charging load, and do not fully consider the spatiotemporal logical relationship between the driving distance of electric vehicles and the location of charging stations, resulting in the planning scheme being out of touch with actual needs; (5) Lack of coordination: The planning of charging stations and the planning of distribution networks are independent of each other, and the two-way linkage between traffic flow and charging load is not realized, which can easily cause the optimization imbalance of a single network and affect the overall operating efficiency of the system.

[0003] Therefore, there is an urgent need for a multi-stage collaborative planning method for charging stations and power distribution networks that can adapt to the dynamic load growth of electric vehicles, take into account the spatiotemporal matching and the full-cycle economy, better meet the growing charging demand and consider the economics of expanding and building new charging stations, accurately describe the logical relationship between the driving distance of electric vehicles and the location of charging stations, and thus plan the location of charging stations more rationally to solve the above problems. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a multi-stage collaborative planning method for charging stations and power distribution networks to address the increasing load on electric vehicles, including: Data initialization: Collect basic road network data, electric vehicle parameters, charging station parameters, total number of motor vehicles in the region, and historical traffic flow data for electric vehicle travel ranges. Establish a real-time data update mechanism to obtain real-time vehicle speeds, congestion indices, number of vehicles queuing at charging stations, demand flow data for electric vehicle travel ranges during specific time periods, and real-time electricity prices. Set key parameters; the electric vehicle travel range includes the starting point, charging station, and destination; key parameters include BPR function parameters, unit time cost, unit energy cost, discount rate, equipment lifespan, and voltage deviation limits. Electric vehicle ownership and charging load prediction in multiple stages: Based on electric vehicle penetration data, the electric vehicle ownership is divided into three stages according to time: basic layout period, rapid expansion period and optimization and improvement period. The electric vehicle ownership prediction model for each stage is used to output the electric vehicle ownership in each region. Combined with dynamic traffic flow results, the charging load of each node of the distribution network at different times is calculated to establish the mapping relationship between traffic flow, charging load and distribution network operation. Charging station site selection optimization: Based on the driving range constraints of electric vehicles, a charging station location model is constructed, range logic constraints are set, charging stations are abstracted as virtual road nodes, and a set of candidate charging stations that meet the range requirements is selected through the dynamic evolution formula of the state of charge. Dynamic traffic flow path allocation: Based on the dynamic user equilibrium model, a dynamic path allocation model is constructed to calculate the generalized cost of comprehensive travel time, energy consumption cost, charging cost and charging waiting time. An improved Logit model is used to solve the path selection probability. According to the flow conservation and road segment capacity constraints, the electric vehicle flow of the road segment is allocated. The constraints are verified and the spatiotemporal coupled travel chain from the starting point through the charging station to the destination is dynamically updated. Verification of the coupled operation of the transportation network and the distribution network: Establish the coupling matrix of the traffic segments and the distribution network nodes, use the Distflow linearization method to calculate the power flow of the distribution network, verify the power balance of the distribution network, the voltage range of the nodes, the capacity of the traffic network segments and the range constraints of electric vehicles, so as to enable the bidirectional linkage between traffic flow and charging load; Multi-stage planning and decision-making for charging stations: With the goal of minimizing the total annual cost within the planning period, the objective function is constructed by integrating the construction cost and operating cost of charging stations. Facility investment and construction constraints and coupling network constraints are set. Based on the electric vehicle ownership, candidate sites and coupling verification results at each stage, the decision is made on the site selection for new charging stations and the expansion scale of existing charging stations at each stage. The optimal planning scheme is output through iterative solution.

[0005] Based on the above technical solution, the present invention can be further improved as follows.

[0006] Furthermore, the electric vehicle ownership prediction model for the corresponding stage is used to output the regional electric vehicle ownership for each stage, including: During the basic layout phase, the penetration rate is set at the beginning of the planning year. The planned starting annual growth rate is , No. The annual policy intensity factor is , No. Annual penetration rate The policy impact coefficient is Then the first Annual penetration rate: ; During the rapid expansion phase, the measured permeability at the end of the first stage is assumed to be... The development coefficient is The amount of ash action is The basic penetration rate is predicted using the grey GM(1,1) model, as follows: ; Traffic flow coupling correction is applied to the basic permeability prediction results, assuming the corrected permeability is... , For the first The total traffic flow within the annual electric vehicle driving range is higher than the planned starting year. To plan the total traffic flow within the driving range of electric vehicles in the initial year, Let the traffic flow connection number be: ; During the optimization and improvement period, the penetration rate saturation growth trend is predicted using the Logistic model, the distribution network carrying capacity is corrected using the distribution network carrying capacity model, and the V2G interaction is corrected using the V2G interaction model. The predicted penetration rate saturation growth trend is assumed to be... , This represents the upper limit of the regional saturated permeability. This represents the measured permeability at the end of the second phase. The growth rate coefficient during the saturation period. If the third phase begins in a certain year, then: ; Let the base value of the stock be The revised distribution network carrying capacity is , For the first Maximum voltage deviation at distribution network nodes in a year For distribution network voltage deviation limits, Assuming the distribution network constraint weights, then: .

[0007] Let the retained quantity after V2G interaction correction be . , For the first Annual V2G discharge penetration rate Let V2G be the correction factor, then: .

[0008] Furthermore, range logic constraints are set, including: no charging stations are set at the starting point and the destination of the electric vehicle; the initial charge of the electric vehicle is sufficient to support driving to the next road node or charging station location; the charge of the electric vehicle is greater than or equal to the expected charge of the electric vehicle after leaving the last node or charging station location after reaching the destination; and the distance between two charging stations is less than the maximum range of the electric vehicle.

[0009] Furthermore, based on the dynamic user equilibrium model, a dynamic route allocation model is constructed to calculate the generalized cost of comprehensive travel time, energy consumption cost, charging cost, and charging waiting time, including: set up Cost per unit driving time for electric vehicles For electric vehicles during the time period on the route Total travel time on the road , For path The set of included road segments For electric vehicles during the time period on the route Average waiting time at charging stations The number of vehicles in the queue. The number of charging stations. , For time period Electric vehicles on the road section Energy consumption costs, The unit charging cost for electric vehicles, For electric vehicles on the path The charging capacity of the charging station. , For path Variables including charging stations, if ,but and ,but: .

[0010] Furthermore, the power flow constraints of the distribution network after linearization using the Distflow linearization method include node active power constraints, reactive power constraints, voltage constraints, and power balance constraints, and the distribution network operates in a radial pattern.

[0011] Furthermore, an improved Logit model is used to solve for the path selection probability, including: assuming for Electric vehicle route selection during time period The probability, The dispersion parameter is used to characterize the user's sensitivity to differences in path costs. Larger users tend to choose the path with the lowest cost. For time period path The generalized cost of electric vehicles, It is an imaginary number. For time period The set of effective driving paths for internal electric vehicles. Let the set of valid paths within the driving range of the electric vehicle be: .

[0012] Furthermore, with the goal of minimizing the total equivalent annual cost within the planning period, an objective function is constructed by integrating the construction and operating costs of charging stations, including: assuming the total equivalent annual cost within the planning period is... , This is the isochronous value coefficient. For the discount rate, This refers to the equipment's lifespan in years. , For the construction cost of charging stations, Let the annual operating cost of the charging station be: .

[0013] Furthermore, a dynamic route allocation model is constructed to calculate the generalized cost of comprehensive driving time, energy consumption cost, charging cost, and charging waiting time. This includes: using a road segment dynamic driving time model to calculate the driving time of electric vehicles with the objectives of minimizing driving time, minimizing charging cost, and minimizing charging waiting time; using a dynamic energy consumption cost model to calculate charging cost; and using a route generalized cost model to calculate the route generalized cost of electric vehicles.

[0014] Furthermore, the construction cost of a charging station includes the purchase of fast charging piles and slow charging piles, site rental, and power access engineering costs; the operating cost of a charging station includes traffic congestion costs, charging and discharging costs, and load fluctuation costs.

[0015] Furthermore, the constraints include facility construction constraints and coupled network constraints; facility construction constraints include the maximum number of devices per station, timing constraints, and charging / discharging power range constraints; coupled network constraints include the correlation constraints between phased station construction decisions and distribution network nodes, traffic network flow conservation, road segment capacity constraints, distribution network power flow balance, and voltage constraints.

[0016] The beneficial effects of this invention are: (1) In view of the dynamic growth characteristics of electric vehicle quantity and charging load, this invention proposes a three-stage stock prediction model and charging station construction and expansion decision-making in the basic layout period, rapid expansion period and optimization and improvement period. By using the equal annual value coefficient to coordinate the construction and operation costs of the whole cycle, the total construction-operation cost of the whole cycle is optimized, which solves the problem that the existing static single-stage planning cannot adapt to long-term load growth. (2) Based on the precise location model of electric vehicle range constraints, this invention constructs a charging station location model that integrates the dynamic evolution of electric vehicle charge state. Through the full-path range logic constraint including the starting point, charging station and the end point, it accurately describes the spatiotemporal logical relationship between driving distance and charging station location, and abstracts the charging station as a virtual road node to optimize the layout, effectively solving the problem of insufficient driving range caused by the disconnect between the distribution of stations and the actual range demand of electric vehicles in traditional planning. (3) This invention proposes a dynamic user equilibrium traffic flow path allocation model. It constructs a dynamic path allocation model with DUE theory as the core, integrates the generalized cost function of travel time, energy consumption cost, charging cost and waiting time, uses an improved Logit model to calculate the path selection probability, and dynamically updates the spatiotemporal coupling travel chain of the starting point, charging station and destination by combining real-time traffic and charging station data. This solves the defect that the traditional static model cannot respond to real-time changes in the road network and ensures that the traffic flow allocation is consistent with the actual operating state. (4) This invention addresses the technical problems of existing static single-stage planning that cannot adapt to the long-term dynamic growth of electric vehicle numbers and charging load, the disconnect between planning schemes and actual needs, and the fragmentation of construction and operation cost control. Through the deep integration of the three-stage electric vehicle ownership prediction model with the collaborative decision-making of new construction and expansion of charging stations, combined with the annual value cost accounting, the planning scheme is accurately matched with the growth characteristics of electric vehicles in different stages, achieving the technical effect of optimal total construction and operation costs within the planning period. This is conducive to the early layout of charging facilities and avoids supply and demand contradictions caused by planning lag. At the same time, through the coordinated planning of transportation and power grid, the risk of imbalance in single network optimization is effectively avoided, and the dynamic electricity demand in the energy transition is adapted. (5) This invention addresses the innovative site selection optimization based on the integration of range constraints and predicted spatiotemporal characteristics. It solves the technical problems of insufficient driving range caused by the disconnect between the spatial distribution of charging stations and the actual range requirements of electric vehicles in traditional charging station planning, and the inability of the station layout to adapt to the spatiotemporal distribution changes of electric vehicles. By combining the full-path range constraints of the state of charge with the spatiotemporal distribution patterns obtained from the prediction of the number of electric vehicles, the optimal location of the charging station is accurately located. This achieves a high degree of matching between the station layout and the driving distance, range capability and future spatiotemporal requirements of electric vehicles, improving the convenience of electric vehicles and user experience. At the same time, it makes the charging station layout more forward-looking and practical, and reduces resource waste. (6) This invention proposes a dynamic traffic flow allocation based on DUE theory and predicted load linkage, which solves the technical problems that traditional static path allocation models cannot respond to changes in real-time road conditions and charging station queuing status, and are difficult to connect with predicted charging load, resulting in a disconnect between traffic flow allocation and actual operating status and future demand. By applying DUE theory and the improved Logit model, and combining the predicted electric vehicle ownership results with real-time data to dynamically update the travel chain, the consistency between traffic flow allocation and real-time road network operating status is achieved. At the same time, it accurately matches the predicted charging load demand, provides users with the most cost-effective path and charging node selection, reduces the overall cost of travel and charging, provides accurate spatiotemporal mapping data of charging load for the power distribution network, ensures the power flow balance and voltage safety of the power distribution network, and improves the overall operating efficiency and stability of the traffic and power grid coupling system. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the multi-stage collaborative planning method for charging stations and distribution networks for electric vehicle load growth provided in Embodiment 1 of the present invention. Figure 2 This is a schematic diagram illustrating the driving range logic of an electric vehicle. Figure 3 A trend chart showing the number of electric vehicles and total charging power; Figure 4 A graph showing the trend of the number of charging facilities in different years; Figure 5 This is a trend chart showing the annual cost structure. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] As an example, see the attached document. Figure 1 As shown, to solve the above-mentioned technical problems, this embodiment provides a multi-stage collaborative planning method for charging stations and distribution networks to address the growth of electric vehicle loads, including: Data initialization: Collect basic road network data, electric vehicle parameters, charging station parameters, total number of motor vehicles in the region, and historical traffic flow data for electric vehicle travel ranges. Establish a real-time data update mechanism to obtain real-time vehicle speeds, congestion indices, number of vehicles queuing at charging stations, demand flow data for electric vehicle travel ranges during specific time periods, and real-time electricity prices. Set key parameters; the electric vehicle travel range includes the starting point, charging station, and destination; key parameters include BPR function parameters, unit time cost, unit energy cost, discount rate, equipment lifespan, and voltage deviation limits. Electric vehicle ownership and charging load prediction in multiple stages: Based on electric vehicle penetration data, the electric vehicle ownership is divided into three stages according to time: basic layout period, rapid expansion period and optimization and improvement period. The electric vehicle ownership prediction model for each stage is used to output the electric vehicle ownership in each region. Combined with dynamic traffic flow results, the charging load of each node of the distribution network at different times is calculated to establish the mapping relationship between traffic flow, charging load and distribution network operation. Charging station site selection optimization: Based on the driving range constraints of electric vehicles, a charging station location model is constructed, range logic constraints are set, charging stations are abstracted as virtual road nodes, and a set of candidate charging stations that meet the range requirements is selected through the dynamic evolution formula of the state of charge. Dynamic traffic flow path allocation: Based on the dynamic user equilibrium model, a dynamic path allocation model is constructed to calculate the generalized cost of comprehensive travel time, energy consumption cost, charging cost and charging waiting time. An improved Logit model is used to solve the path selection probability. According to the flow conservation and road segment capacity constraints, the electric vehicle flow of the road segment is allocated. The constraints are verified and the spatiotemporal coupled travel chain from the starting point through the charging station to the destination is dynamically updated. Verification of the coupled operation of the transportation network and the distribution network: Establish the coupling matrix of the traffic segments and the distribution network nodes, use the Distflow linearization method to calculate the power flow of the distribution network, verify the power balance of the distribution network, the voltage range of the nodes, the capacity of the traffic network segments and the range constraints of electric vehicles, so as to enable the bidirectional linkage between traffic flow and charging load; Multi-stage planning and decision-making for charging stations: With the goal of minimizing the total annual cost within the planning period, the objective function is constructed by integrating the construction cost and operating cost of charging stations. Facility investment and construction constraints and coupling network constraints are set. Based on the electric vehicle ownership, candidate sites and coupling verification results at each stage, the decision is made on the site selection for new charging stations and the expansion scale of existing charging stations at each stage. The optimal planning scheme is output through iterative solution.

[0020] To accurately describe the logical relationship between the driving distance of electric vehicles and the location of charging stations, and to more rationally plan the locations of charging stations, this invention proposes an electric vehicle charging station location model based on the driving range constraints of electric vehicles. (See attached...) Figure 2The diagram illustrates the driving range logic of an electric vehicle, using an Origin-Destination Pair (OD Pair) and a 5-node road network. The OD Pair starts at node 1 and ends at node 5, with only one 5-node path between them. The road lengths are L1, L2, L3, and L4, and the road numbers are as follows: , , and The driving route is , , and .

[0021] Assume the initial charge of the electric vehicle starting from node 1 is... The estimated charge after reaching the destination and subsequently leaving node 5 is If the electric vehicle has a charge of [missing information] after a full charge, , The corresponding longest driving range is Furthermore, to better reflect real-world conditions, the charging stations are considered for placement in the middle of the road, and their location can be viewed as a virtual road node. Before selecting the charging station location, three assumptions are made regarding the range constraints of electric vehicles: (1) No charging stations are set up at the starting and ending points of electric vehicles; (2) For the initial charge of electric vehicles It should be able to sustain driving to the next road node or charging station; considering range anxiety, there should be some remaining battery power when reaching the destination. ; (3) The distance between two charging stations is less than the maximum driving range of the electric vehicle. .

[0022] Based on the above analysis, the corresponding... Figure 1 Logical constraints on the driving distance of electric vehicles under a 5-node road: To ensure that electric vehicles can reach the first candidate charging station from the starting node. ; the charge state at the constraint starting point is greater than 0, ;set up Indicates the first One road, and This indicates the distance between two candidate charging station locations. For roads The decision variable for whether to build a new charging station, and the state-of-charge constraint for electric vehicles traveling between charging stations, are expressed as follows: ; ; The state of charge of the electric vehicle when it reaches the destination node is higher than , is represented as: .

[0023] Extending the relationship between electric vehicle range constraints and candidate charging station locations to a more general scenario: Let... Indicates the starting point of transportation. Indicates the destination of a transportation trip. Indicates link At transportation hubs, for each OD pair The set of connected paths is , path , This represents a path that is related to the origin of the OD (Original Design Environment). The set of links, Is OD to The set of all paths between, and paths between specific OD pairs. The driving range logical constraint is expressed as a constraint on the electric vehicle's state of charge at the starting point, as follows: ; Computational Link State of charge of the electric vehicle: ; Calculation path Uplink State of charge of the electric vehicle: ; Constraining electric vehicles on the path k any link on At this point, the state of charge of the electric vehicle is always non-negative: ; Constraining the state of charge of the electric vehicle as it approaches the destination node: ; Using regional electric vehicle penetration rate as the core prediction, and considering multi-stage planning of charging stations, the system is divided into three stages: basic layout, rapid expansion, and optimization. Combining policy factors, traffic flow data, power distribution network constraints, and V2G dynamic adjustments, the system outputs the electric vehicle ownership for each time stage and region. For the first The annual number of electric vehicles is the core output of the multi-stage electric vehicle number growth prediction model. For the first The total number of motor vehicles in the region each year is taken from the annual statistics of the public security traffic management department. For the first Annual electric vehicle penetration rate, key predictor variables at each stage This is the annual data correction factor, used to correct statistical bias, calculated as: previous year's measured stock / previous year's predicted stock. (Region number...) The core formula for calculating the annual number of electric vehicles is: .

[0024] Optionally, the electric vehicle ownership prediction model for the corresponding stage can be used to output the regional electric vehicle ownership for each stage, including: During the basic layout phase, the penetration rate is set at the beginning of the planning year. The planned starting annual growth rate is , No. The annual policy intensity factor is , No. Annual penetration rate The policy impact coefficient is Then the first Annual penetration rate: ; During the rapid expansion phase, the measured permeability at the end of the first stage is assumed to be... The development coefficient is The amount of ash action is The basic penetration rate is predicted using the grey GM(1,1) model, as follows: ; Traffic flow coupling correction is applied to the basic permeability prediction results, assuming the corrected permeability is... , For the first The total traffic flow within the annual electric vehicle driving range is higher than the planned starting year. To plan the total traffic flow within the driving range of electric vehicles in the initial year, Let the traffic flow connection number be: ; During the optimization and improvement period, the penetration rate saturation growth trend is predicted using the Logistic model, the distribution network carrying capacity is corrected using the distribution network carrying capacity model, and the V2G interaction is corrected using the V2G interaction model. The predicted penetration rate saturation growth trend is assumed to be... , This represents the upper limit of the regional saturated permeability. This represents the measured permeability at the end of the second phase. The growth rate coefficient during the saturation period. If the third phase begins in a certain year, then: ; Let the base value of the stock be The revised distribution network carrying capacity is , For the first Maximum voltage deviation at distribution network nodes in a year For distribution network voltage deviation limits, Assuming the distribution network constraint weights, then: .

[0025] Let the retained quantity after V2G interaction correction be . , For the first Annual V2G discharge penetration rate Let V2G be the correction factor, then: .

[0026] This model constructs a three-stage charging station planning and prediction framework: the basic layout stage, the rapid expansion stage, and the optimization and improvement stage. In the basic stage, linear regression and policy correction are used to capture the steady growth characteristics driven by policies. In the expansion stage, gray GM(1,1) is coupled with traffic flow to accurately depict the spatiotemporal distribution differences during the demand explosion stage. In the optimization stage, Logistic saturation growth and bidirectional correction of distribution network and V2G are used to respond to the core needs of slowing penetration rate and vehicle-to-grid interaction. The model iteratively updates the correction coefficients through measured data to achieve a step-by-step improvement in prediction accuracy.

[0027] Optionally, range logic constraints can be set, including: no charging stations are set at the starting point and the destination of the electric vehicle; the initial charge of the electric vehicle is sufficient to support driving to the next road node or charging station location; the charge of the electric vehicle is greater than or equal to the expected charge of the electric vehicle after leaving the last node or charging station location after reaching the destination; and the distance between two charging stations is less than the maximum range of the electric vehicle.

[0028] Optionally, based on a dynamic user equilibrium model, a dynamic route allocation model is constructed to calculate the generalized cost of comprehensive travel time, energy consumption cost, charging cost, and charging waiting time, including: set up Cost per unit driving time for electric vehicles For electric vehicles during the time period on the route Total travel time on the road , For path The set of included road segments For electric vehicles during the time period on the route Average waiting time at charging stations The number of vehicles in the queue. The number of charging stations. , For time period Electric vehicles on the road section Energy consumption costs, The unit charging cost for electric vehicles, For electric vehicles on the path The charging capacity of the charging station. , For path Variables including charging stations, if ,but and ,but: .

[0029] Optionally, the power flow constraints of the distribution network after linearization using the Distflow linearization method include node active power constraints, reactive power constraints, voltage constraints, and power balance constraints, and the distribution network operates in a radial pattern.

[0030] From the perspective of the distribution network physical model, the distribution network in the planning scenario mainly operates radially, with relatively significant line impedance and dispersed load distribution. Traditional nonlinear power flow equations rely on iterative solutions and are sensitive to initial values, making it difficult to directly integrate into the multivariable and multi-constraint planning model framework. Distflow linearization, by ignoring higher-order terms in line power losses, transforms the nonlinear voltage drop relationship into linear constraints (such as the linear relationship between the square of node voltage and line power). While retaining the core physical logic of power flow calculation, it significantly simplifies the mathematical expression of the distribution network power flow model, enabling it to accurately adapt to the operating characteristics of radial distribution networks. Simultaneously, from the perspective of multi-model coupling solution, the charging station planning model needs to couple the real-time data of electric vehicle charging loads from dynamic traffic flow output with the distribution network power flow, while integrating economic objective functions such as multi-stage construction costs and operating costs of charging stations, forming a complex optimization model with multiple constraints and objectives. Distflow... The linearized power flow constraints can be seamlessly integrated with the traffic flow-charging load mapping model and the charging station construction and operation cost model, transforming them into a second-order cone programming form that is easy to solve efficiently. This effectively avoids the convergence difficulties or local optima problems that may occur in multivariate iterations of nonlinear models, ensuring that the globally optimal charging station planning scheme can be quickly solved under the premise of satisfying constraints such as power balance and voltage safety of the distribution network, thus guaranteeing the feasibility and efficiency of coordinated planning of traffic and power grid.

[0031] set up For nodes Transmit to node Net active power, For nodes Transmit to node Net reactive power, For nodes voltage, For nodes To node The resistance of the circuit, For nodes To node The reactance of the line, For nodes The basic load of the connection, For the first Electric vehicle charging power, For nodes The minimum value of the voltage, For nodes If the maximum voltage is: ; ; ; ; .

[0032] The charging station is connected to the distribution network node, and the distribution network supplies power to the charging station. After a traffic accident, the electric vehicle charging load is included in the distribution network load, and the voltage situation of the distribution network node is analyzed. Simultaneously, the distribution network operates radially, and DistFlow is used for power flow calculations. Optionally, an improved Logit model can be used to solve for the path selection probability, including: setting for Electric vehicle route selection during time period The probability, The dispersion parameter is used to characterize the user's sensitivity to differences in path costs. Larger users tend to choose the path with the lowest cost. For time period path The generalized cost of electric vehicles, It is an imaginary number. For time period The set of effective driving paths for internal electric vehicles. Let the set of valid paths within the driving range of the electric vehicle be: .

[0033] Optionally, with the goal of minimizing the total annualized cost within the planning period, an objective function is constructed by integrating the construction and operating costs of charging stations, including: assuming the total annualized cost within the planning period is... , This is the isochronous value coefficient. For the discount rate, This refers to the equipment's lifespan in years. , For the construction cost of charging stations, Let the annual operating cost of the charging station be: .

[0034] set up This is a set of candidate locations for fast charging stations. This is a set of candidate locations for slow charging stations. This is a collection of fast charging station models. This is a collection of slow charging station models. For the stage In position The number of fast charging stations to be built For the stage In position The number of slow charging stations to be built Model The unit cost of fast charging stations Model The unit cost and construction cost of a slow charging station are: Traffic congestion costs are , Cost per unit time for The actual travel time of electric vehicles during the specified time period. For the actual travel time of gasoline-powered vehicles, For the free driving time of electric vehicles, Free time for gasoline-powered vehicles. for Demand flow for electric vehicles during specific time periods for The demand flow of gasoline vehicles during a given time period, and the charging and discharging cost are , For load fluctuation costs, for Time-of-use electricity pricing For charging stations exist The total charging and discharging power during the time period, with charging being positive and discharging being negative. Unit fluctuation cost, For charging stations The average load is: ; ; ; ; ; ; .

[0035] Optionally, a dynamic route allocation model is constructed to calculate the generalized cost of comprehensive driving time, energy consumption cost, charging cost, and charging waiting time. This includes: using a road segment dynamic driving time model to calculate the driving time of electric vehicles, using a dynamic energy consumption cost model to calculate the charging cost, and using a route generalized cost model to calculate the route generalized cost of electric vehicles, with the goal of minimizing driving time, minimizing charging cost, and minimizing charging waiting time.

[0036] Based on the theory of Dynamic User Equilibrium (DUE), this method dynamically adjusts the path selection behavior of electric vehicles through real-time traffic data, establishing a spatiotemporally coupled travel chain between the origin, charging station, and destination. This addresses the shortcomings of traditional static traffic flow path assignment models, which cannot respond to changes in road conditions, and satisfies the requirements of any origin-destination pair. At any given time, the generalized cost of all used paths is equal to the minimum path cost, and the generalized cost of unused paths is not less than the minimum path cost, thus ensuring that traffic flow allocation is consistent with the actual road network operation.

[0037] Route selection criteria: Electric vehicles make route decisions with the goal of minimizing driving time, charging cost, and charging waiting time; Real-time data input: It is assumed that real-time vehicle speed, congestion index, and number of vehicles queuing at charging stations can be obtained for each road segment of the road network, with a data update cycle of 15 minutes; Range constraint premise: The driving path of electric vehicles must meet the range requirements in the charging station location model, that is, the distance between any two adjacent charging stations is less than or equal to the maximum range of the electric vehicle.

[0038] Travel time calculation: The travel time of electric vehicles on a road segment is dynamically adjusted according to the traffic flow of electric vehicles, and the travel time of electric vehicles during the time period is calculated. The travel time on the road is set as follows For road section Free passage time (travel time without congestion) is determined by basic road network data; and For the parameters of the BPR function (BPR travel time function), usually It is 0.15. A score of 4 reflects the degree to which road congestion affects travel time. For time period Section Electric vehicle traffic, For road section If the maximum traffic capacity is such that the flow of electric vehicles does not exceed the traffic capacity, then: ; Dynamic energy consumption cost calculation: Establish an energy consumption cost model for electric vehicles, converting electricity consumption into monetary costs, unifying the target dimension for path selection, and setting... For time period Section Real-time energy consumption of electric vehicles , , and These are energy consumption parameters, typically taken as -8.54, 0.3, 298.3, and 0.0024, respectively. The unit energy cost of electric vehicles For time period Section The real-time driving speed of electric vehicles is obtained from real-time traffic data. For road section If the length is such that the real-time energy consumption of the electric vehicle is: ; The energy cost of electric vehicles is: .

[0039] Path generalized cost calculation: Let Cost per unit driving time for electric vehicles For time period EVs in the path Total travel time on the road , For electric vehicles during the time period on the route Average waiting time at charging stations , The number of charging stations. For time period Electric vehicles on the road section Energy consumption costs, The unit charging cost for electric vehicles, For electric vehicles on the path The charging capacity of the charging station. , For path The variables containing 0 and 1 of the charging station, if ,but and .

[0040] Optionally, the construction cost of a charging station includes the purchase of fast charging piles and slow charging piles, site rental, and power access engineering costs; the operating cost of a charging station includes traffic congestion costs, charging and discharging costs, and load fluctuation costs.

[0041] Optionally, the constraints include facility construction constraints and coupled network constraints; facility construction constraints include the maximum number of devices per station, timing constraints, and charging / discharging power range constraints; coupled network constraints include the correlation constraints between phased station construction decisions and distribution network nodes, traffic network flow conservation, road segment capacity constraints, distribution network power flow balance, and voltage constraints.

[0042] set up For time period OD to path Electric vehicle traffic, for Time period OD The total demand flow of electric vehicles is obtained from regional traffic demand forecasting, and the OD (Original Demand) conservation of demand flow is as follows: ;set up For time period Section Traffic flow of electric vehicles For road section Path association matrix, if road segment Belongs to OD pair path , but Otherwise Traffic flow allocation for road sections is as follows: Path flow is related to selection probability as follows: ; Dynamic traffic flow allocation constraints: To ensure that the traffic flow of electric vehicles does not exceed the road segment's capacity and to avoid exacerbating congestion, the following constraints apply: The electric vehicle range constraint is based on the electric vehicle charging station location model to ensure that the electric vehicle's battery does not run out while driving. ,in, Nonnegativity constraints require that parameters such as flow rate and probability be nonnegative. .

[0043] Multi-stage constraints include facility deployment constraints, power constraints, and coupled network constraints. Let... This represents the maximum number of devices per station. This refers to a fast charging station. This refers to a slow charging station. To correspond to the total number of fast charging stations to be built within the planned time period, To correspond to the total number of slow charging stations to be built within the planned time period, Decision variables for charging station site selection The number of charging devices configured for a single charging station, with quantity constraints expressed as follows: ; Timing constraints are expressed as: ; set up Minimum charging and discharging power of slow charging stations This represents the maximum charging and discharging power of the slow charging station. This refers to the minimum charging and discharging power of the fast charging station. The maximum charging and discharging power of the fast charging station is expressed as follows: ; ; set up For the stage Section The decision to build the website For the stage Distribution network nodes The decision to build the website Let be the coupling matrix. Traffic network constraints include flow conservation and segment capacity constraints in the dynamic traffic flow model; distribution network constraints include power flow balance and voltage constraints in the coupled model. Therefore, the connection constraints are expressed as: .

[0044] Model Validation: A traffic network and an IEEE 69-node distribution network coupled together were used as the test system to verify the effectiveness of the electric vehicle charging station planning model proposed in this patent. (See attached diagram) Figure 3 The trend chart showing the number of electric vehicles and total charging power indicates that the number of electric vehicles and total charging power are growing in tandem, verifying the accuracy of the vehicle ownership-charging load mapping relationship in the multi-stage planning, and demonstrating that the power supply capacity can dynamically adapt to the growing demand for electric vehicles. (Attached) Figure 4 The chart showing the trend of charging facility quantity changes over different years indicates that the number of charging facilities (stations / charging piles) increases with the planning period and expands rapidly in the later stages, aligning with the phased rhythm of basic layout followed by rapid expansion. The growth rate of charging piles is higher than that of charging stations, but it has not exceeded the upper limit of single-station equipment, demonstrating that the facility configuration is both adaptable and compliant. (Attached) Figure 5 The annual cost structure trend chart shows that the total annual cost steadily increases with the expansion of the facility scale, with land costs accounting for a significantly larger proportion in the later stages. The cost structure is synergistically linked with facility construction and load growth demand, with no obvious resource redundancy (wasteful vacancy of charging facilities) or investment imbalance. The transportation network consists of 24 nodes and 76 roads; the IEEE 69-node distribution network has 69 nodes and 68 lines.

[0045] Based on the spatiotemporal coupled travel chain logic, charging nodes that meet the optimal balance between range constraints and cost are selected: Candidate charging stations are selected from the set of candidate sites output by the electric vehicle charging station location model. In the process, select those located at OD pairs Charging stations along the effective path form a path. Candidate charging node set The goal of charging node selection is to choose the charging node that minimizes the total generalized cost of the origin, charging node, and destination for any electric vehicle travel path. for , of which For time period Electric vehicles from the starting point To charging node Generalized cost, For time period Electric vehicles from charging nodes To the finish line The generalized cost, in yuan; For time period Electric vehicles at charging nodes The charging cost (including waiting cost). The travel chain is dynamically updated to reflect the charging cost at a certain charging node. Number of vehicles in queue Increased If the travel time is extended due to rising traffic or road congestion, the generalized cost and probability of the route will be recalculated, and the electric vehicle travel chain will be dynamically updated to ensure that the route selection always meets the dynamic user equilibrium conditions.

[0046] The iterative method for solving the dynamic traffic flow path assignment model includes the following steps: (1) Initialization: Input basic road network data, including road segment lengths Free passage time , traffic capacity Electric vehicle parameters (including energy consumption parameters) (and initial charge state) and power station parameters (including the number of charging piles and charging cost); (2) Input real-time data: obtain time period Real-time traffic data (road segment speed) Charging station queue data and OD to demand flow ; (3) Calculate the travel time and energy cost of the road segment: calculate in sequence , and ; (4) Calculate the total cost of the path: ; (5) Calculate the path selection probability: Calculate ; (6) Distribute traffic flow to road sections: Calculate and ; (7) Constraint check: Check whether the constraint conditions are met. If they are met, proceed to the next step. If they are not met, adjust the path selection probability (e.g., increase the cost weight of the violation path) and recalculate. (8) Generate travel chains: Select the optimal charging points and generate a time-space integrated travel chain for electric vehicles; (9) Selected Generation Update: Time Period After the solution is completed, Repeat steps (2)-(8) until all time periods of the day have been solved; (10) Output results: Output electric vehicle traffic flow, path selection probability, electric vehicle travel chain and charging node selection results for each time period and road segment, providing input data for the coupled operation model of power distribution network and transportation network.

[0047] set up For electric vehicle charging rate, To connect with distribution network nodes A set of coupled traffic segments. for Time period and road section Electric vehicle traffic, For distribution network nodes exist The electric vehicle charging load during a given time period, represented by the coupled operation model of the power distribution network and transportation network, is as follows: .

[0048] for Time Periods and Road Sections Traffic flow of fuel-powered vehicles For road section The maximum traffic capacity, or road segment capacity constraint, is expressed as: The range constraint of electric vehicles is expressed as .make sure Time period and road section The electric vehicle on board has a non-negative state of charge to avoid power outages during operation.

[0049] This invention proposes a multi-stage electric vehicle (EV) ownership prediction-driven collaborative planning model for charging stations and power distribution networks. Addressing the dynamic growth characteristics of EV numbers and charging load, it proposes a three-stage ownership prediction model encompassing the basic layout phase, rapid expansion phase, and optimization and improvement phase, along with decisions on new construction and expansion of charging stations. By using an equal-annual coefficient to coordinate the entire lifecycle construction and operation costs, it achieves optimal total construction-operation costs throughout the entire lifecycle, solving the problem that existing static single-stage planning cannot adapt to long-term load growth.

[0050] This invention is based on a precise location model constrained by electric vehicle range. It constructs a charging station location model that integrates the dynamic evolution of the electric vehicle's state of charge. By including the full-path range logic constraint of the starting point, charging station, and destination, it accurately depicts the spatiotemporal logical relationship between driving distance and charging station location. It abstracts charging stations as virtual road nodes for optimized layout, effectively solving the problem of insufficient driving range caused by the disconnect between the distribution of stations and the actual range requirements of electric vehicles in traditional planning.

[0051] This invention proposes a dynamic user equilibrium traffic flow path allocation model. Based on DUE theory, a dynamic path allocation model is constructed, integrating a generalized cost function of travel time, energy consumption cost, charging cost, and waiting time. An improved Logit model is used to calculate the path selection probability. By combining real-time traffic (vehicle speed and congestion) and charging station (queuing vehicles) data, the spatiotemporal coupled travel chain of the starting point, charging station, and destination is dynamically updated. This solves the problem that traditional static models cannot respond to real-time changes in the road network and ensures that traffic flow allocation is consistent with the actual operating state.

[0052] This invention addresses the challenges of multi-stage electric vehicle (EV) ownership prediction-driven collaborative planning. It solves the technical problems of existing static single-stage planning failing to adapt to the long-term dynamic growth of EV numbers and charging load, the disconnect between planning schemes and actual needs, and fragmented construction and operation cost control. By deeply integrating a three-stage ownership prediction model with collaborative decision-making on new and expanded charging stations, and combining it with equivalent annual cost accounting, the invention achieves precise matching between planning schemes and the growth characteristics of EVs at different stages. This results in the optimal total construction and operation cost within the planning period, facilitating the early deployment of charging facilities and avoiding supply-demand imbalances caused by planning delays. Furthermore, through coordinated planning of transportation and the power grid, it effectively mitigates the risk of imbalance in single-network optimization and adapts to the dynamic electricity demand during energy transition.

[0053] This invention addresses the innovative site selection optimization approach that integrates range constraints and predicted spatiotemporal characteristics. It solves the technical problems of insufficient driving range caused by the disconnect between the spatial distribution of charging stations and the actual range requirements of electric vehicles in traditional charging station planning, and the inability of station layouts to adapt to changes in the spatiotemporal distribution of electric vehicles. By combining the full-path range constraint based on the state-of-charge (SOC) with the spatiotemporal distribution patterns derived from the prediction of electric vehicle ownership, the optimal location of charging stations (abstracted as virtual road nodes) is precisely determined. This achieves a high degree of matching between station layout and electric vehicle driving distance, range capability, and future spatiotemporal needs, improving the convenience and user experience of electric vehicles. Simultaneously, it makes charging station layout more forward-looking and practical, reducing resource waste.

[0054] This invention proposes a dynamic traffic flow allocation based on DUE theory and predicted load linkage. It solves the technical problems of traditional static route allocation models being unable to respond to real-time road network conditions (such as congestion) and changes in charging station queuing status, and having difficulty connecting with predicted charging load, leading to a disconnect between traffic flow allocation and actual operating status and future demand. By applying DUE theory and an improved Logit model, and combining electric vehicle ownership prediction results with real-time data to dynamically update travel chains, consistency between traffic flow allocation and real-time road network operating status is achieved. At the same time, it accurately matches predicted charging load demand, providing users with cost-optimal route and charging node selection, reducing the overall cost of travel and charging, providing accurate spatiotemporal mapping data of charging load for the power distribution network, ensuring power flow balance and voltage safety of the power distribution network, and improving the overall operating efficiency and stability of the traffic and power grid coupled system.

[0055] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for multi-stage coordinated planning of charging stations and power distribution networks in view of growth of electric vehicle loads, characterized in that, Comprise: Data initialization: Collect road network basic data, electric vehicle parameters, charging station basic parameters, regional total motor vehicle inventory and historical traffic flow of electric vehicle driving range, establish real-time data updating mechanism, obtain real-time road segment speed, congestion index, charging station queue vehicle number, time period electric vehicle driving range demand flow and real-time electricity price, set key parameters; Electric vehicle driving range includes driving starting point, charging station and driving end point; Key parameters include BPR function parameters, unit time cost, unit energy consumption cost, discount rate, equipment life and voltage deviation limit; Electric vehicle multi-stage inventory and charging load prediction: According to the electric vehicle penetration rate data, the electric vehicle inventory is divided into three stages according to time: basic layout period, rapid expansion period and optimization and improvement period, the corresponding stage electric vehicle inventory prediction model is adopted to output the regional electric vehicle inventory of each stage, and the dynamic traffic flow result is combined to calculate the charging load of each node of the distribution network at different time periods, and the mapping relationship among traffic flow, charging load and distribution network operation is established; Charging station site optimization: Based on the constraint of electric vehicle driving range, a charging station location model is constructed, the driving logic constraint is set, the charging station is abstracted as a virtual road node, and the candidate charging station set meeting the driving requirements is selected through the state of charge dynamic evolution formula; Dynamic traffic flow path allocation: Based on the dynamic user equilibrium model, a dynamic path allocation model is constructed, the generalized cost of comprehensive driving time, energy consumption cost, charging cost and charging waiting time is calculated, the improved Logit model is used to solve the path selection probability, the electric vehicle flow of road section is allocated according to the flow conservation and road section capacity constraint, the constraint condition is checked and the time and space coupled travel chain from the starting point to the charging station to the end point is dynamically updated; Traffic network and distribution network coupled operation verification: The coupling matrix of traffic section and distribution network node is established, the Distflow linearization method is used for distribution network power flow calculation, the power balance of distribution network, node voltage range and traffic section capacity and electric vehicle driving range constraint are checked, so that the two-way linkage of traffic flow and charging load is realized; Multi-stage planning and decision of charging station: Taking the minimum total equivalent annual cost in the planning period as the target, the construction cost and operation cost of charging station are integrated to construct the objective function, the facility construction constraint and coupled network constraint are set, based on the electric vehicle inventory, candidate station and coupling verification result of each stage, the new site selection and expansion scale of existing charging station of each stage are decided, and the optimal planning scheme is output by iteration solution.

2. The method of claim 1, wherein, The corresponding stage electric vehicle inventory prediction model is adopted to output the regional electric vehicle inventory of each stage, including: The basic layout period is set as the planning starting year, the penetration rate is , the growth rate in the planning starting year is , the policy intensity factor in the year is , the penetration rate in the year is , and the policy influence coefficient is . Therefore, the penetration rate in the year is: ； The measured permeability at the end of the first stage is , the development coefficient is , and the ash content is . The base permeability is predicted by using the grey GM(1, 1) model, which is expressed as ； The traffic flow coupling correction is made to the basic permeability prediction result, and the corrected permeability is set as , is the total flow of the electric vehicle driving range in the first year, is the increment of the total flow of the electric vehicle driving range compared with the planning starting year, is the total flow of the electric vehicle driving range in the planning starting year, is the traffic flow contact number, and then: ； Optimization and improvement stage, through the Logistic model prediction of the saturation growth trend of penetration rate, using the distribution network carrying capacity model to correct the distribution network carrying capacity, using the V2G interaction model to correct the V2G interaction, and the saturation growth trend prediction result of penetration rate is , is the upper limit of regional saturation penetration rate, is the measured penetration rate at the end of the second stage, is the saturation growth rate coefficient, is the third stage starting year, then: ； The amount of insurance is based on the value of , the modified distribution network carrying capacity is , The maximum voltage deviation of the distribution network node in the first year, The voltage deviation limit of the distribution network, The distribution network constraint weight, then: 。 Let the corrected stock of V2G interaction be , the first year V2G discharge penetration rate, V2G correction coefficient, then: 。 3. The method of claim 1, wherein, Set driving logic constraint, including: the starting point and the end point of the electric vehicle are not set with charging station; The starting electric quantity of the electric vehicle can support driving to the next road node or charging station location, and the electric quantity of the electric vehicle is greater than or equal to the expected electric quantity of the electric vehicle after reaching the last node or charging station location; The distance between two charging stations is less than the maximum driving range of the electric vehicle.

4. The method of claim 1, wherein, Based on the dynamic user equilibrium model, a dynamic path assignment model is constructed to calculate the generalized cost of the comprehensive travel time, energy consumption cost, charging cost and charging waiting time, including: Let be the cost of the electric vehicle per unit travel time, be the total travel time of the electric vehicle on the path in the time period, , be the set of road segments contained in the path , be the average waiting time of the electric vehicle at the charging stations on the path in the time period, be the number of vehicles in the queue, be the number of charging posts, , be the energy consumption cost of the electric vehicle on the road segment in the time period, be the cost of the electric vehicle per unit charging, be the charging amount of the electric vehicle at the charging stations on the path in the time period, , be the variable of the charging station contained in the path , if , and , 。 5. The method of claim 1, wherein, The power flow constraints of the linearized distribution network by the Distflow linearization method include active power constraints, reactive power constraints, voltage constraints and power balance constraints, and the distribution network operates in a radial manner.

6. The method of claim 1, wherein, An improved Logit model is used to solve for path selection probabilities, including: setting for Electric vehicle route selection during time period The probability, The dispersion parameter is used to characterize the user's sensitivity to differences in path costs. Larger users tend to choose the path with the lowest cost. For time period path The generalized cost of electric vehicles, It is an imaginary number. For time period The set of effective driving paths for internal electric vehicles. Let the set of valid paths within the driving range of the electric vehicle be: 。 7. The method of claim 1, wherein, With the minimum total equivalent annual cost in the planning period as the goal, the construction cost and operation cost of charging stations are integrated to build the objective function, including: setting the total equivalent annual cost in the planning period as , is the equivalent annual coefficient, is the discount rate, is the equipment life, , is the construction cost of charging stations, is the annual operation cost of charging stations, then: 。 8.The method of claim 1, wherein, A dynamic path assignment model is constructed to calculate the generalized cost of the comprehensive travel time, energy consumption cost, charging cost and charging waiting time, including: taking the minimum travel time, the lowest charging cost and the shortest charging waiting time as the target, using the dynamic travel time model to calculate the travel time of the electric vehicle, using the dynamic energy consumption cost model to calculate the charging cost, and using the path generalized cost model to calculate the path generalized cost of the electric vehicle. 9.The method of claim 1, wherein, The construction cost of the charging station includes the equipment procurement of fast charging piles and slow charging piles, site rental and power access engineering cost; the operation cost of the charging station includes traffic congestion cost, charging and discharging cost and load fluctuation cost.

10. The method of claim 1, wherein, The constraint conditions include facility construction constraints and coupled network constraints; the facility construction constraints include single station maximum equipment quantity limit, time sequence constraint and charging and discharging power range constraint; the coupled network constraints include stage station construction decision and associated constraint of distribution network node, traffic network flow conservation, road section capacity constraint, distribution network power flow balance and voltage constraint.

Images