A high-speed road network planning method and system based on an extended traffic network flow allocation model

Abstract

The present invention discloses a highway network planning method and system based on an extended traffic network flow allocation model, including: constructing a traffic flow allocation model based on spatio-temporal paths, combining the user equilibrium mechanism, considering the game of electric vehicle users in terms of departure time, route selection and charging mode, and capturing the interdependence between charging behavior and route selection; performing highway charging network planning based on a robust optimization framework, with the optimization objective of maximizing the profit of private investors; adopting a novel reconstruction method based on untruncated cuts to solve the problem of embedding the inner-layer binary decision variables of robust optimization. The method of the present invention can effectively optimize the construction of highway charging infrastructure and achieve the optimal allocation of charging load.

Description

Technical Field

[0001] The present invention relates to a high - speed road network planning method based on an extended traffic network flow distribution model, belonging to the technical fields of electric vehicle charging management and smart grid. Background Art

[0002] With the increasing global emphasis on the "low - carbon" initiative and the breakthroughs in electric vehicle technology, especially in fields such as batteries, electric motors, and electronic control systems, the penetration rate of electric vehicles in the global transportation system has been continuously rising. This trend has put forward higher requirements for charging infrastructure. Especially on inter - city roads, the charging demands of electric vehicles are becoming more complex and diverse.

[0003] During inter - city driving, the energy consumption of electric vehicles is significantly higher than that in urban driving. Due to the increased air resistance and reduced motor energy efficiency during high - speed driving, the unit energy consumption of electric vehicles during long - distance driving has increased substantially. Therefore, a complete inter - city charging infrastructure has become a key factor to ensure that electric vehicle users can complete long - distance trips smoothly.

[0004] In addition, the construction of inter - city charging infrastructure not only concerns the satisfaction of electric vehicle charging demands but also directly affects the stability of the power grid. The spatio - temporal distribution differences of charging loads in different regions, especially the concentrated charging demands of electric vehicles at specific times, may pose great pressure on the power grid and even lead to the risk of power grid load overload or congestion. The spatial and temporal flexibility of electric vehicle charging behavior provides an opportunity to relieve the power grid pressure. If the charging time and location can be reasonably scheduled, the charging load can be effectively dispersed, thus reducing the power grid burden.

[0005] However, the existing electric vehicle charging load scheduling methods have limitations. Traditional modeling methods usually rely on accurate travel data of electric vehicle users and a single charging mode. However, in practical applications, the charging load is affected by many factors, including users' travel plans, charging station layouts, charging price fluctuations, and traffic flows. Especially on inter - city sections, due to the changes in high - speed driving and traffic flows, the spatio - temporal dynamic characteristics of charging demands and power grid loads have not been effectively considered, resulting in the failure of existing scheduling methods to fully reflect the diversification and dynamics of electric vehicle charging demands, affecting the flexibility potential of electric vehicle charging load regulation and the expected benefits of planning operators. Summary of the Invention

[0006] The technical problem to be solved by the present invention is that in the prior art, the distribution of electric vehicle charging loads lacks effective regulation, resulting in unbalanced power grid loads, affecting the stable operation of the power grid. At the same time, the synergistic effects among charging station operators, electric vehicle users, and the power grid have not been fully considered, and the optimal distribution of charging loads cannot be achieved.

[0007] To solve the above technical problems, the present invention provides a high-speed road network planning method based on an extended traffic network flow distribution model, including:

[0008] Step S1: Collect power grid dispatching information, collect traffic network topology and electric vehicle travel demand data for each time period, and collect investment prices of charging infrastructures with different powers;

[0009] Step S2: Based on the collected traffic network topology and electric vehicle travel demand data for each time period obtained in Step S1, construct a flow distribution model for the extended traffic network, including a main problem and a sub-problem. The main problem realizes the optimal traffic flow distribution under a given spatio-temporal path, and the goal of the sub-problem is to determine the optimal spatio-temporal path of electric vehicle users according to the current traffic conditions in different EOD pairs;

[0010] Step S3: After solving the model in Step S2 to obtain the optimal traffic flow distribution, calculate the charging loads of fast charging and slow charging by using the extended spatio-temporal path flow and the corresponding charging amounts;

[0011] Step S4: Based on the power grid dispatching information obtained in Step S1, construct a demand response model that integrates the interaction among the power grid, charging station operators, and electric vehicle users;

[0012] Step S5: Based on the infrastructure investment price obtained in Step S1, construct an objective function for the charging station operator CSO, and the objective function includes planning cost, charging cost, and demand response cost. The demand response cost is determined according to whether the charging load in Step S4 meets the power grid regulation requirements;

[0013] Step S6: Use matrices to represent the charging infrastructure planning model and solve it. Among them, the charging infrastructure planning model includes planning variables, state variables in the dispatching stage, and constraint conditions;

[0014] The planning variables include binary variables used to represent whether to build a charging station in the planning cost, and integer variables used to build fast charging piles, slow charging piles, and vehicle-to-grid interaction charging piles;

[0015] The state variables in the dispatching stage include charging load, charging price in the charging cost, and extended spatio-temporal path flow;

[0016] The constraint conditions are used to represent the coupling constraints between the planning stage and the dispatching stage.

[0017] The aforesaid method for planning a high-speed road network based on an extended traffic network flow distribution model. In step S2, the master problem transmits the real-time state of the traffic network to the sub-problem. The sub-problem uses the shortest path algorithm to find the optimal spatio-temporal path and returns the optimal spatio-temporal path as an alternative travel plan to the master problem. After iteration, the optimal traffic flow distribution of the ETAP-UE model is finally obtained.

[0018] The aforesaid method for planning a high-speed road network based on an extended traffic network flow distribution model. In step S2, the extended traffic network is a three-dimensional directed graph, denoted as , where the set represents extended nodes, including actual traffic network nodes and virtual slow charging nodes at different time periods; represents the road sections between nodes;

[0019] Each node is represented as , where represents traffic network nodes and virtual slow charging nodes, represents the time section index and set, represents the road section index and set, which is composed of a pair of traffic network nodes ; is the road section set. In the extended traffic network, each electric vehicle user travels between the starting point and the ending point, forming an extended origin-destination pair EOD, , represents the set of all EOD pairs. Each EOD pair is represented as , where and represent the starting point and ending point positions, and represent the arrival time and the expected departure time respectively. For electric vehicles, the feasible spatio-temporal path is indexed by , and the corresponding spatial path is indexed by ; represents the spatio-temporal path set, represents the two-dimensional spatial path index, represents the two-dimensional spatial path set;

[0020] The master problem is expressed as follows:

[0021] (1)

[0022] (2)

[0023] (3)

[0024] (4)

[0025] (5)

[0026] (6)

[0027] (7)

[0028] (8)

[0029] (9)

[0030] (10)

[0031] In the formula, and respectively represent the electric vehicle flow on the road section at time and the number of electric vehicles queuing at the charging station ; The functions and respectively represent the travel time on the road section and the queuing time at the charging station ; The parameters and respectively represent the free flow travel time on the road section and the inherent service time of the charging station ; and are respectively the capacities of the road section and the charging station ; The parameter represents the travel demand of the OD pair ; is the conversion factor from time to monetary cost; represents the total travel cost of the OD pair on the path ; represents the minimum travel cost of the OD pair ; The indicators and respectively represent the relationship between the path and the road section or the charging station; represents the mapping relationship between the spatio-temporal path and the spatial path ; represents the charging station queuing time constant; represents the set of fast charging stations; represents the spatio-temporal path travel cost;​ Represents the time cost caused by changing travel plans. Represents the set of slow charging stations. Represents complementary slack constraints. Represents the connection between the two-dimensional space path and the charging stations. Represents the connection between the two-dimensional space path and the traffic network segments;

[0032] The sub-problem is represented as follows:

[0033] (11)

[0034] (12)

[0035] (13)

[0036] (14)

[0037] (15)

[0038] (16)

[0039] (17)

[0040] (18)

[0041] (19)

[0042] (20)

[0043] In the formula, the variables and respectively represent the charging prices of slow charging and fast charging stations at time moment. represents the remaining battery power of the electric vehicle at location and time moment. For fast charging, and respectively represent the charging amount and the maximum power. For slow charging, and and respectively represent the charging amount, the maximum power and the minimum power. is the slack variable of the unselected path. is the distance between location and . is the energy consumption coefficient per unit distance. is used for large The constant in the method, the remaining battery power remains higher than the preset threshold throughout the journey , represents the spatio-temporal path selection 0-1 variable, represents the section flow calculated by the master problem, represents the charging station queue volume calculated by the master problem, is the auxiliary slack variable for the unselected charging station, is when traveling the electric vehicle power at point is when traveling the electric vehicle power at point represents the fast charging amount when traveling, is the travel time, is the originally scheduled travel time, is the next moment the electric vehicle power at point is the next moment, represents the expanded node, is the initial power, represents the power at time 0, represents point the electric vehicle power at time represents the minimum SOC caused by the anxiety of electric vehicle users, represents the maximum power of the electric vehicle battery.

[0044] The aforementioned high-speed road network planning method based on the expanded traffic network flow allocation model, in step S3, calculates the charging loads of fast charging and slow charging, and the expression is as follows:

[0045] (21)

[0046] (22)

[0047] In the formula, and respectively represent the charging load amounts in the fast charging and slow charging modes; and respectively represent the fast charging and slow charging amounts corresponding to the expanded spatio-temporal path .

[0048] The aforementioned high-speed road network planning method based on the expanded traffic network flow allocation model, in step S4, the demand response model is as shown in formula (23):

[0049] (23).

[0050] The aforementioned method for planning a high-speed road network based on an extended traffic network flow allocation model, in step S5, the objective function of the power station operator CSO is as shown in equation (24).

[0051] (24)

[0052] In the formula, , and respectively represent the construction costs of charging stations, fast charging piles, and slow charging piles; , and respectively represent the numbers of charging stations, proposed fast charging piles, and slow charging piles to be constructed. and respectively represent the charging prices set by the CSO for fast charging and slow charging. and respectively represent the electricity selling prices provided by the power grid. and respectively represent the demand response costs of fast charging stations and slow charging stations.

[0053] The CSO planning constraints are as follows:

[0054] (25)

[0055] (26)

[0056] (27)

[0057] (28)

[0058] In the formula, represents the planning budget of the CSO.

[0059] The aforementioned method for planning a high-speed road network based on an extended traffic network flow allocation model performs robust optimization on the CSO objective function and planning constraints constructed in step S5 to achieve the goal of maximizing the CSO profit under the influence of uncertain factors.

[0060] The aforementioned method for planning a high-speed road network based on an extended traffic network flow allocation model, the robust optimization is expressed as:

[0061] (29)

[0062] (30)

[0063] (31)

[0064] (32)

[0065] In the formula, the overline and the underline respectively represent the upper and lower limits of the relevant variables.

[0066] The aforementioned method for planning a high-speed road network based on an extended traffic network flow distribution model is characterized in that, in step S6, a matrix is used to represent a high-speed highway charging infrastructure planning model, which is expressed as:

[0067] (33)

[0068] In the formula, the first-stage variable is a planning variable; the second-stage variable is the state variable in the scheduling stage, including the charging loads and under fast charging and slow charging modes, the set charging prices and under fast charging and slow charging modes, and the extended spatio-temporal path flow ;

[0069] The constraint condition is , which represents the coupling constraint between the planning stage and the scheduling stage, corresponding to constraint conditions (27)-(28);

[0070] The set represents the feasible region of the planning variable, corresponding to constraint conditions (25)-(26);

[0071] The set represents the feasible region of the scheduling variable, with equations (1)-(10), (21)-(22) as constraint conditions;

[0072] The set represents the uncertainty set of the parameters, with equations (29)-(32) as constraint conditions;

[0073] represents the uncertain parameter, represents the feasible region of the second-stage variable affected by the first-stage variable , , , are respectively the parameter one, parameter two, and parameter three for the compact expression of the constraint;

[0074] Using the convex relaxation technique, the feasible region of the original problem (33) exhibits the dual property, and the inner-layer maximization expression is transformed into the minimization form ;

[0075] denote the convex hull of is an auxiliary variable denote the -dimensional non-negative real vector space, i.e., each element in the vector is a non-negative real number. The superscript denotes the transpose of the vector; denote the transpose of the vector;

[0076] The dual property is where denote the extreme point solution of

[0077] Using the disjunctive cut constraint (38), the original robust problem (33) is reformulated as an RO-MP framework problem, and then the RO-MP framework problem is input into the solver for solution;

[0078] The RO-MP framework expression is:

[0079] (34)

[0080] (35)

[0081] (36)

[0082] (37)

[0083] (38)

[0084] (39)

[0085] In the formula, the disjunctive cut , , is the extreme point set;

[0086] For let where the part after the vertical bar is the condition and the part before the vertical bar is the range. Among them, is the -th unit vector, -th represents the -th term of The symbol denotes the number of elements in the set, denotes the set excluding the set ;

[0087] , , , are all compact constraint matrices obtained by combining the preset constraint conditions, is an auxiliary convergence decision variable, represents the set of unprofitable cuts, represents the th extreme point.

[0088] A computer system includes a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps of the method as described above.

[0089] The beneficial effects achieved by the present invention: In the method of the present invention, a traffic flow allocation model is designed by combining a spatio-temporal path traffic assignment model with a time-expanded user equilibrium mechanism. The traffic flow allocation model includes multi-dimensional decision-making behaviors of electric vehicle users in departure time selection, path planning decision-making, and charging demand response, providing a basis for the layout optimization of highway charging infrastructure. By constructing a robust optimization decision model with the goal of maximizing the profit of charging station operators, aiming at the solution problem caused by the binary decision variables in the inner layer of the model, a model reconstruction method based on unprofitable cut sets is used, and through mathematical derivation, the continuous conversion of the mixed integer problem is realized, significantly improving the calculation efficiency and achieving the optimal allocation of charging load. BRIEF DESCRIPTION OF THE DRAWINGS

[0090] Figure 1 is a flowchart of a highway network planning method based on an extended traffic network flow allocation model in Embodiment 1 of the present invention;

[0091] Figure 2 is a schematic diagram of an extended traffic network in Embodiment 1 of the present invention;

[0092] Figure 3 is a traffic flow heat map in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0093] The technical solution of the present invention will be further described below with reference to the drawings.

[0094] As Figure 1 shown, this embodiment provides a highway network planning method based on an extended traffic network flow allocation model, including the following steps:

[0095] Step S1: Collect power grid dispatching information, collect traffic network topology and electric vehicle travel demand data in each time period, and collect investment prices of charging infrastructure with different powers; the dispatching information includes historical electricity sales prices, demand response signals, etc.;

[0096] Historical electricity sales prices and demand response signals in power grid dispatching information, which give electricity price fluctuations and power grid load conditions at different times, providing a basis for estimating the operating costs of charging facilities and pricing strategies; traffic network topology and electric vehicle travel demand data at each time period, which give the highway layout, electric vehicle flow, and travel patterns, and are used to identify key regions and time periods with high charging demand; investment price data of charging infrastructure with different powers, which are used to evaluate construction costs and select appropriate equipment types and quantities to maximize economic benefits.

[0097] Step S2: Based on the traffic network topology and electric vehicle travel demand data at each time period obtained in Step S1, construct a traffic flow allocation model for the expanded traffic network (expand traffic assignment problem with user equilibrium, ETAP-UE), including a master problem and a sub-problem. The master problem realizes the optimal traffic flow distribution given a set of spatio-temporal paths. The goal of the sub-problem is to determine the optimal spatio-temporal paths of electric vehicle users according to the current traffic conditions in different EOD pairs. The master problem transfers the real-time state of the traffic network to the sub-problem. The sub-problem uses the shortest path algorithm to find the optimal spatio-temporal paths and returns the optimal spatio-temporal paths to the master problem as alternative travel plans. After iteration, the optimal traffic flow distribution of the ETAP-UE model is finally obtained.

[0098] Since the spatio-temporal paths are finite based on the departure time and network topology, the feasible travel options for electric vehicle users are also finite. Therefore, the iteration process can ensure convergence.

[0099] The expanded transportation network (ETN) is a three-dimensional directed graph, denoted as , where the set represents the expanded nodes, including the actual traffic network nodes and virtual slow charging nodes at different time periods; represents the road segments between nodes, which are spatio-temporal road segments, including time road segments and space road segments. The time road segment is a path that spans different time periods of the same physical node or different virtual slow charging nodes, and the space road segment is a path that spans different locations within the same time period. As shown in Figure 2 , the virtual slow charging nodes encountered by electric vehicle users are actually located at the same physical location.

[0100] Each node is represented as , where represents the traffic network nodes and virtual slow charging nodes, represents the time section index and set. Therefore, any spatio-temporal path can be represented as , where and are elements of, in addition, represents the road segment index and set, which consists of a pair of traffic network nodes . is the set of road segments. In the ETN (Extended Traffic Network), each electric vehicle user travels between the origin and destination, forming an expanded origin-destination pair EOD (expanded origin destination pair, EOD). , where represents the set of all EOD pairs. Different from the traditional TAP-UE model, the definitions of the origin and destination in the ETN have space, time, and energy-related attributes, including location, arrival time, initial battery level, and expected departure time. Each EOD pair is represented as , where and represent the origin and destination locations, and represent the arrival time and expected departure time respectively. For electric vehicles, the feasible spatio-temporal path is indexed by , and the corresponding spatial path is indexed by . represents the set of spatio-temporal paths, represents the two-dimensional spatial path index, represents the set of two-dimensional spatial paths;

[0101] The main problem is expressed as follows:

[0102] (1)

[0103] (2)

[0104] (3)

[0105] (4)

[0106] (5)

[0107] (6)

[0108] (7)

[0109] (8)

[0110] (9)

[0111] (10)

[0112] Wherein, and respectively represent the electric vehicle flow at time on section and the number of electric vehicles queuing at the charging station ; The functions and respectively represent the travel time on section and the queuing time at the charging station ; The parameters and respectively represent the free flow travel time on section and the inherent service time of the charging station ; and are respectively the capacities of section and the charging station ; The parameter represents the travel demand of OD pair ; is the conversion factor from time to monetary cost (i.e., the time-money conversion coefficient), represents the total travel cost of OD pair on path , including travel time, queuing time, charging cost, etc., represents the minimum travel cost of OD pair ; The indicators and respectively represent the relationships between path and the section or the charging station; If the flow on section passed by path is , then = 1, otherwise it is 0; represents the mapping relationship between the spatio-temporal path and the spatial path ; represents the charging station queuing time constant, represents the set of fast charging stations, represents the spatio-temporal path travel cost, represents the time cost caused by changing the travel plan, represents the set of slow charging stations, represents the complementary slack constraint, represents the connection between the two-dimensional spatial path and the charging station, represents the connection between the two-dimensional spatial path and the traffic network section.

[0113] The constraint equation (1) represents the relationship between the spatio-temporal path and the travel demand, while the constraint equation (2) ensures that the traffic flow is non-negative. The constraint equation (3) combines the three-dimensional spatio-temporal path that includes travel time, charging decision, and spatial path selection with the two-dimensional spatial path that includes time period and spatial information and establishes a one-to-one correspondence through auxiliary variables between . The constraint equations (4) and (5) are used to calculate the travel time and the queuing time at the charging station respectively, while the constraint equations (6) and (7) are used to calculate the travel cost. The constraint equation (8) represents the equilibrium state after the user behavior stabilizes, and the constraint equations (9) and (10) relate the spatial path to the link flow and the queuing vehicle flow at the charging station.

[0114] The sub-problem is expressed as follows:

[0115] (11)

[0116] (12)

[0117] (13)

[0118] (14)

[0119] (15)

[0120] (16)

[0121] (17)

[0122] (18)

[0123] (19)

[0124] (20)

[0125] In the formula, the variables and represent the charging prices of the slow charging station and the fast charging station at time respectively. represents the remaining battery power of the electric vehicle at location at time . For fast charging, , represent the charging amount and the maximum power respectively. For slow charging, , and represent the charging amount, the maximum power and the minimum power respectively. The slow charging station also supports discharging through vehicle-to-grid (V2G) technology, and the minimum power of discharging is set to -7 kW. is the slack variable of the unselected path. When the path is selected, = 1; represents the initial battery charge of the electric vehicle. is the distance between the positions and . is the energy consumption coefficient per unit distance. is a very large constant used in the large method. Considering the anxiety of users, the remaining battery charge is kept higher than the preset threshold throughout the journey. represents the spatio-temporal path selection 0-1 variable. represents the section flow calculated by the master problem. represents the charging station queue volume calculated by the master problem. is the auxiliary slack variable of the unselected charging station. is the electric vehicle power at the time of the trip. is the electric vehicle power at the time of the trip. represents the fast charging amount during the trip. is the travel time. is the originally scheduled travel time. is the electric vehicle power at the next moment . is the next moment. represents the expanded node. is the initial charge. represents the charge at time 0. represents the electric vehicle power at the point and the moment. represents the minimum SOC caused by the anxiety of electric vehicle users. represents the maximum battery charge of the electric vehicle.

[0126] The constraint equation (11) is the objective function of the sub-problem, which optimizes the spatio-temporal path selection by minimizing the travel cost of electric vehicle users. The travel cost includes travel time, charging station queuing time, energy costs of slow charging at home and fast charging on highways, and also includes additional costs caused by delays or changes in travel plans. The constraint equations (12) and (13) handle the spatial path conversion within the same time period and the temporal path conversion at the same location respectively, and the two together constitute the mapping relationship of the spatio-temporal path. It is assumed that the travel time of the electric vehicle does not exceed 1 hour. If this time limit is exceeded, two OD pairs are constructed by introducing virtual intermediate nodes to shorten the journey. The constraint equations (14) and (15) limit the fast charging and slow charging amounts respectively, and the constraint equations (16) and (17) represent the slack variables of unselected paths and paths without charging. The constraint equation (18) represents the initial battery power of the electric vehicle, the constraint equation (19) represents the user anxiety problem caused by excessive travel time, and the constraint (20) represents the additional cost caused by the user changing the plan.

[0127] The traffic flow allocation model for the extended transportation network can capture the coupling of fast charging and slow charging, comprehensively consider the mutual influence of the two, and optimize the charging facility configuration. At the same time, the model can accurately reflect the actual travel time of users, reasonably arrange the location and capacity of charging facilities, form an equilibrium charging network layout, improve resource utilization efficiency, reduce waiting time, enhance the user experience, and achieve the coordination and unity of charging network planning and user needs.

[0128] Step S3: After solving the model in Step S2 to obtain the optimal traffic flow distribution, use the extended spatio-temporal path flow and the corresponding charging amounts to calculate the charging loads of fast charging and slow charging. The expressions are as follows:

[0129] (21)

[0130] (22)

[0131] In the formula, and respectively represent the charging load amounts in the fast charging and slow charging modes; and respectively represent the fast charging and slow charging amounts corresponding to the extended spatio-temporal path . When the auxiliary variable or takes a value of 1, it indicates that this path passes through the location at time and selects the corresponding charging mode (fast charging or slow charging). The extended spatio-temporal path refers to the spatial path extended in the time dimension, so it is a spatio-temporal path.

[0132] Step S4: Based on the power grid dispatching information obtained in Step S1, construct a demand response model that integrates the interactions among the power grid, charging station operators, and electric vehicle users. According to the charging load status obtained in Step S3, when the charging load meets the power grid regulation requirements, the power grid gives corresponding rewards to the charging station operators; if the charging load exceeds the regulation upper limit, corresponding penalties are given to the charging station operators; when the charging load is lower than the regulation lower limit, it has no impact on the charging station operators. The specific constraint conditions are shown in Equation (23):

[0133] (23)

[0134] In the formula, the upper and lower limits of the demand response for fast and slow charging loads are represented by 、 、 and respectively. The corresponding demand response prices for fast and slow charging are represented by and respectively. The above-set limits and prices constitute the basis for implementing rewards and penalties for the CSO charging load in the demand response mechanism.

[0135] The demand response information is an uncertain parameter. The demand response information parameters for the charging station operators are unknown and can only be revealed during the intraday dispatching process. Therefore, in Step S6, a robust optimization method is used to optimize the demand response parameters.

[0136] Step S5: Based on the infrastructure investment prices for different charging powers obtained in Step S1, construct the objective function and planning constraints of the charging station operator CSO, with the goal of maximizing the profit of the charging station operator CSO;

[0137] To simultaneously meet the travel and charging needs of electric vehicle users and the peak shaving requirements of the power grid, the objective function includes the planning cost, charging cost, and demand response cost. Among them, the demand response cost is related to the demand response model established in Step S4, that is, the demand response cost of the CSO is determined according to whether the charging load in Step S4 meets the power grid regulation requirements; the objective function of the power station operator CSO is shown in Equation (24),

[0138] (24)

[0139] In the formula, 、 and represent the construction costs of charging stations, fast charging piles, and slow charging piles respectively; 、 and represent the quantities of charging stations, proposed fast charging piles, and slow charging piles to be built respectively, and respectively represent the charging prices set by the CSO for fast charging and slow charging, and respectively represent the electricity selling prices provided by the power grid; and respectively represent the demand response costs of fast charging stations and slow charging stations.

[0140] The CSO planning constraints are as follows:

[0141] (25)

[0142] (26)

[0143] (27)

[0144] (28)

[0145] In the formula, represents the planning budget of the CSO, indicating that all planning costs do not exceed the financial capacity of the CSO.

[0146] Constraint (26) means that the charging station needs to be built first and then the fast charging pile. Constraints (27) and (28) are both used to limit that the charging station needs to meet the charging load demand of users.

[0147] The robust optimization method is used to handle the electricity price fluctuations and the uncertainty of demand response DR information faced by the CSO during operation. The uncertain factors are unknown information for the CSO and only the basic range can be inferred based on historical data. Therefore, an interval uncertainty set is constructed to characterize the uncertain factors. The CSO objective function and planning constraints constructed in step S5 are robustly optimized to achieve the goal of maximizing the CSO profit under the influence of uncertain factors, while meeting the travel and charging needs of electric vehicle users and the peak shaving requirements of the power grid. The robust optimization is expressed as:

[0148] (29)

[0149] (30)

[0150] (31)

[0151] (32)

[0152] In the formula, the overline and underline respectively represent the upper and lower limits of the relevant variables. The upper and lower limit thresholds of the demand response price are set based on the historical maximum and minimum values, and the same applies to the power grid electricity selling price.

[0153] Step S6: Further simplify the expression and represent the highway charging infrastructure planning model using matrices as follows:

[0154] (33)

[0155] In the formula, the first-stage variable is all the planning variables, including the binary variables for building charging stations, i.e., 0-1 variables, and the integer variables for building fast charging piles, slow charging piles, and vehicle-to-grid (V2G) charging piles; the second-stage variable is all the state variables related to the dispatching stage, including the charging loads and in fast charging and slow charging modes, the set charging prices and in the set fast charging and slow charging modes, and the extended spatio-temporal path flow and so on.

[0156] The constraint condition is , which represents the coupling constraint between the planning stage and the dispatching stage, corresponding to constraint conditions (27)-(28);

[0157] The set represents the feasible region of the planning variables, corresponding to constraint conditions (25)-(26);

[0158] The set represents the feasible region of the dispatching variables, with equations (1)-(10), (21)-(22) as the constraint conditions;

[0159] The set represents the uncertainty set of the parameters, with equations (29)-(32) as the constraint conditions;

[0160] represents the uncertain parameter, represents the feasible region of the second-stage variables affected by the first-stage variable , , , are the parameter one, parameter two, and parameter three for the compact expression of the constraint respectively.

[0161] Due to the existence of non-linear factors such as variable multiplication or conditional statements in the above constraint conditions and objective function, the proposed electric vehicle highway charging infrastructure planning model is transformed into a mixed-integer linear programming problem through piecewise linearization, the big-M method, and SOS constraints, which belongs to the prior art and will not be elaborated here.

[0162] Since piecewise linearization, the big-M method, and SOS constraints all introduce binary variables, making it difficult for traditional robust optimization dual methods to solve the primal problem in Equation (33) of Step S6. To solve this problem, the present invention proposes a No-Good-Cut-Based method, which uses convex relaxation technology to make the feasible region (Feasible Region, the range of values that variables can take) of the original problem in Equation (33) exhibit dual properties, and transforms the inner maximization expression into a minimization form ;

[0163] denotes the convex hull of is an auxiliary variable, denotes the n-dimensional non-negative real vector space, that is, each element in the vector is a non-negative real number, and the superscript denotes the transpose of the vector, that is, transposing a column vector into a row vector for matrix multiplication operations.

[0164] The dual property is where denotes the extreme point solution of

[0165] To improve the accuracy of the robust optimization solution, using the No-Good-Cut constraint in Equation (38), the original robust problem in Equation (33) is reformulated as an RO-MP framework problem, and then the reformulated RO-MP framework problem is input into a solver for solution.

[0166] The RO-MP framework expression is:

[0167] (34)

[0168] (35)

[0169] (36)

[0170] (37)

[0171] (38)

[0172] (39)

[0173] In the formula, the No-Good-Cut , , is the extreme point set;

[0174] For , let , the part after the vertical line is the condition, and the part before the vertical line is the range, where is the th unit vector, represents the th item, the symbol represents the number of elements in the set, represents the set excluding the set ;

[0175] , , , are all compact constraint matrices combined by preset constraint conditions, is an auxiliary convergence decision variable, represents the set of non-good cuts, represents the th extreme point.

[0176] is the first-stage variable in the RO-MP framework problem, which has been relaxed. The RO-MP framework problem is iteratively solved by the branch and bound algorithm to obtain an exact solution.

[0177] This embodiment uses a test system based on an actual traffic network. The traffic network consists of 31 nodes and 114 links. Node #31 and Node #15 are designated as the starting point and the slow charging station candidate node, and Nodes #30, #13, and #3 are designated as the destinations. All other nodes are regarded as potential fast charging station candidate nodes. 45,122 electric vehicles are simulated. The driving range of the electric vehicles is 400 kilometers, and the initial SoC is set to 10%.

[0178] Figure 3 (a) shows the traffic flow distribution results of the traditional TAP-UE model and the proposed traffic flow assignment model using extended spatio-temporal paths at 9 am. By introducing the game of electric vehicle users in travel time, electric vehicle users choose to depart early or late, and avoid severe congestion during the morning rush hour by adjusting the departure time. Compared with the situation shown in Figure 3 (a), Figure 3 (b) shows that the congestion situation has been significantly alleviated because some electric vehicle users choose to travel during periods of low traffic demand.

[0179] Embodiment 2

[0180] A computer system includes a memory, a processor, and a computer program stored on the memory. The processor executes the computer program to implement the steps of the method described in Embodiment 1.

[0181] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device produce a means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0182] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory produce a manufacture including instruction means that implement the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0183] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0184] The above is only the preferred embodiment of the present invention. It should be noted that for those of ordinary skill in the art of the present technology, without departing from the technical principle of the present invention, several improvements and deformations can be made, and these improvements and deformations should also be regarded as the protection scope of the present invention.

Claims

1. A highway network planning method based on an extended traffic network flow allocation model, characterized in that Including: Step S1: Collect power grid dispatching information, collect traffic network topology and electric vehicle travel demand data for each period, and collect investment prices of charging infrastructures with different powers; Step S2: Based on the collected traffic network topology and electric vehicle travel demand data for each period obtained in Step S1, construct a traffic flow distribution model for the extended traffic network, including a main problem and a sub-problem. The main problem realizes the optimal traffic flow distribution under a given spatio-temporal path, and the goal of the sub-problem is to determine the optimal spatio-temporal path of electric vehicle users according to the current traffic conditions among different EOD pairs; Step S3: After solving the model in Step S2 to obtain the optimal traffic flow distribution, calculate the charging loads of fast charging and slow charging by using the extended spatio-temporal path flow and the corresponding charging amounts; Step S4: Based on the power grid dispatching information obtained in Step S1, construct a demand response model that integrates the interactions among the power grid, charging station operators, and electric vehicle users; Step S5: Based on the infrastructure investment prices obtained in Step S1, construct a target function for the charging station operator CSO, where the target function includes planning costs, charging costs, and demand response costs, and the demand response costs are determined according to whether the charging loads in Step S4 meet the power grid regulation requirements; Step S6: Represent the charging infrastructure planning model by using matrices and solve it. Among them, the charging infrastructure planning model includes planning variables, state variables in the dispatching stage, and constraint conditions; The planning variables include binary variables used to represent whether to build a charging station in the planning costs, and integer variables used to build fast charging piles, slow charging piles, and vehicle-grid interaction charging piles; The state variables in the dispatching stage include charging loads, charging prices in the charging costs, and extended spatio-temporal path flows; The constraint conditions are used to represent the coupling constraints between the planning stage and the dispatching stage.

2. The method for planning a high-speed road network based on an extended traffic network flow distribution model according to claim 1, wherein, In Step S2, the main problem transfers the real-time state of the traffic network to the sub-problem. The sub-problem uses the shortest path algorithm to find the optimal spatio-temporal path and returns the optimal spatio-temporal path to the main problem as an alternative travel plan; through iteration, the optimal traffic flow distribution of the ETAP-UE model is finally obtained.

3. A highway network planning method based on an extended traffic network flow distribution model according to claim 2, characterized in that In step S2, the extended traffic network is a three-dimensional directed graph, denoted as , where the set represents the extended nodes, including the actual traffic network nodes and virtual slow charging nodes at different time periods; represents the road segments between the nodes; Each node is represented as , where represents traffic network nodes and virtual slow charging nodes, represents time section indices and sets, represents road section indices and sets, which are composed of a pair of traffic network nodes constituting is the set of road sections. In the extended traffic network, each electric vehicle user travels between a starting point and an ending point, forming an extended origin-destination pair EOD, , represents the set of all EOD pairs, and each EOD pair is represented as , where and represent the starting point and ending point locations, and represent the arrival time and the expected departure time respectively. For electric vehicles, the feasible spatio-temporal path is indexed by , and the corresponding spatial path is indexed by . represents the set of spatio-temporal paths, represents two-dimensional spatial path indices, represents the set of two-dimensional spatial paths; The main problem is expressed as follows: （1） （2） （3） （4） （5） （6） （7） （8） （9） （10） Wherein, and respectively represent the electric vehicle flow on the road section at time and the number of electric vehicles queuing at the charging station ; the functions and and respectively represent the travel time on the road section and the queuing time at the charging station ; the parameters and respectively represent the free flow travel time on the road section and the inherent service time of the charging station ; and are respectively the capacities of the road section and the charging station ; the parameter represents the travel demand of the OD pair ; is the conversion factor from time to monetary cost, represents the total travel cost of the OD pair on the path ; represents the minimum travel cost of the OD pair ; the indicators and respectively represent the relationships between the path and the road section or the charging station; represents the mapping relationship between the spatio-temporal path and the spatial path ; represents the charging station queuing time constant, represents the set of fast charging stations, represents the spatio-temporal path travel cost, represents the time cost caused by changing the travel plan, represents the set of slow charging stations, represents the complementary slack constraint, represents the connection between the two-dimensional spatial path and the charging station, represents the connection between the two-dimensional spatial path and the traffic network road section; The sub-problem is expressed as follows: （11） （12） （13） （14） （15） （16） （17） （18） （19） （20） Wherein, the variables and respectively represent the charging prices of the slow charging station and the fast charging station at time . represents the remaining battery power of the electric vehicle at location at time . For fast charging, and respectively represent the charging amount and the maximum power. For slow charging, , and respectively represent the charging amount, the maximum power and the minimum power. is the slack variable of the unselected path. is the distance between location and . is the energy consumption coefficient per unit distance. is a constant used in the large method. The remaining battery power remains higher than the preset threshold during the entire journey. represents the spatio-temporal path selection 0-1 variable. represents the section flow calculated by the master problem. represents the charging station queue volume calculated by the master problem. is the auxiliary slack variable of the unselected charging station. is the electric vehicle power at point during the journey. is the electric vehicle power at point during the journey. represents the fast charging amount during the journey. is the travel time. is the originally scheduled travel time. is the electric vehicle power at the next moment . is the next moment. represents the expanded node. is the initial power. represents the power at time 0. represents at point the electric vehicle power at the moment. represents the minimum SOC caused by the anxiety of the electric vehicle user. represents the maximum battery power of the electric vehicle.

4. A highway network planning method based on an extended traffic network flow distribution model according to claim 3, characterized in that, In Step S3, the charging loads of fast charging and slow charging are calculated, and the expression is as follows: （21） （22） Wherein, and respectively represent the charging load amounts in fast charging and slow charging modes; and respectively represent the fast charging and slow charging amounts corresponding to the extended space-time path ​ 5. A highway network planning method based on an extended traffic network flow distribution model according to claim 4, characterized in that, In Step S4, the demand response model is as shown in Equation (23): （23）。 6. The high-speed road network planning method based on an extended traffic network flow distribution model according to claim 5, wherein In Step S5, the target function of the power station operator CSO is as shown in Equation (24), （24） Wherein, , and respectively represent the construction costs of charging stations, fast charging piles, and slow charging piles; , and respectively represent the quantities of charging stations, fast charging piles to be built, and slow charging piles to be built, and respectively represent the charging prices set by the CSO for fast charging and slow charging, and respectively represent the electricity selling prices provided by the power grid; and respectively represent the demand response costs of fast charging stations and slow charging stations; The CSO planning constraints are as follows: （25） （26） （27） （28） In the formula, represents the planned budget of the CSO.

7. A highway network planning method based on an extended traffic network flow distribution model according to claim 6, characterized in that Robustly optimize the CSO target function and planning constraints constructed in Step S5 to achieve the goal of maximizing the CSO profit under the influence of uncertain factors.

8. A highway network planning method based on an extended traffic network flow distribution model according to claim 7, characterized in that The robust optimization is expressed as: （29） （30） （31） （32） In the formula, the overline and underline respectively represent the upper and lower limits of the relevant variables.

9. A highway network planning method based on an extended traffic network flow distribution model according to claim 7, characterized in that In Step S6, represent the highway charging infrastructure planning model by using matrices, which is expressed as: （33） In the formula, the first-stage variable is a planning variable; the second-stage variable is the state variable in the scheduling stage, including the charging loads and under fast charging and slow charging modes, the charging prices and set under fast charging and slow charging modes, and the extended spatio-temporal path flow ; The constraint condition is , representing the coupling constraint between the planning stage and the scheduling stage, corresponding to the constraint conditions in equations (27)-(28); Set represents the feasible region of the planning variable, corresponding to the constraint equations (25)-(26); Set represents the feasible region of the scheduling variable, with equations (1)-(10) and equations (21)-(22) as the constraint conditions; Set Represents an uncertain set of parameters, subject to the constraints of equations (29)-(32); Represents an uncertain parameter, Represents the feasible region of the second-stage variable affected by the first-stage variable, , , Are the parameter one, parameter two, and parameter three for the compact representation of constraints respectively; Using convex relaxation techniques, the feasible region of the original problem in Equation (33) exhibits dual properties, and the maximization expression in the inner layer is transformed into a minimization form ; denote the convex hull of is an auxiliary variable denote the vector space of non - negative real numbers of dimension denotes the transpose of the vector; The dual property is , where denotes the extreme point solution of Use the disjunctive cut constraint Equation (38) to reformulate the original robust problem Equation (33) into a RO-MP framework problem, and then input the RO-MP framework problem into the solver for solution; The RO-MP framework expression is: （34） （35） （36） （37） （38） （39） In the formula, there is no good cut , , is the pole set; For , let . Behind the vertical bar is the condition, and in front of the vertical bar is the range, where is the -th unit vector, represents 's -th term, the symbol represents the number of elements in the set, represents the set except for the set ; the remaining part. , , , are all compact constraint matrices merged from preset constraint conditions, is an auxiliary convergence decision variable, represents the set of unprofitable cuts, represents the th extreme point.

10. A computer system, comprising a memory, a processor, and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the steps of the method according to any one of claims 1-9.

Images