A power traffic coupling system control method based on an alternating direction multiplier method

Abstract

The application provides a power traffic coupling system control method based on an alternating direction multiplier method, based on a vehicle-road-station-network collaborative architecture, real-time collection of traffic side data, power side data and environment side data; according to a traffic network, a power distribution network and a charging station virtual link, a traffic network extension network model is constructed, and a generalized path cost function of heterogeneous users is constructed; a Logit discrete selection model is used to output the charging flow distribution of each charging station; based on the traffic side data, the power side data, the environment side data and the charging flow distribution, combined with a second-order cone relaxation algorithm, a power distribution network optimal power flow model is obtained; according to the power distribution network optimal power flow model, the electrical safety constraint and the power distribution network operation state, combined with an improved alternating direction multiplier method, a guiding price of charging station power is dynamically generated; through the above technical scheme, the coordination optimization of traffic flow and power distribution network operation state under multi-agent information interaction is realized.

Description

Technical Field

[0001] This application relates to the field of electric-transportation coupling system management technology, and in particular to a control method for electric-transportation coupling systems based on the alternating direction multiplier method. Background Technology

[0002] With the large-scale deployment of electric vehicles, traditional centralized optimization methods have significant limitations in terms of model solution scale, information sharing requirements, and privacy protection. At the same time, existing technologies mostly assume that charging users are a homogeneous group, failing to consider the differences in travel time, path impedance, and charging cost perception among different types of users, and thus failing to capture the chain dynamic effect of "perceptual heterogeneity - path and charging selection - local over-limitation of distribution network".

[0003] The purpose of the distributed collaborative optimization method for power-transportation coupled systems that considers differences in user travel modes is to address the challenges of information dispersion, privacy protection, and collaborative regulation in systems with multiple independent operators by introducing refined modeling of user perception heterogeneity. By finely characterizing the path-charging joint selection behavior of users with different behavioral preferences and combining it with an improved distributed solution algorithm, the method achieves coordinated optimization of traffic flow and power grid operation while ensuring the privacy of local information on both the transportation and distribution sides. This avoids road congestion and localized grid overload caused by excessive spatial aggregation of electric vehicles. Summary of the Invention

[0004] This application provides a power-transportation coupled system control method based on the alternating direction multiplier method, which solves the limitation of the prior art in accurately capturing the chain dynamic effect of "perception heterogeneity - path and charging selection - local limit exceedance of distribution network". It realizes the accurate characterization of the spatiotemporal evolution law of electric vehicle charging load and provides a more realistic joint selection behavior modeling tool for power-transportation coupled system analysis.

[0005] To achieve the above objectives, embodiments of this application provide a control method for an electric-transit coupled system based on the alternating direction multiplier method, the method comprising: Based on a vehicle-road-station-network collaborative architecture, real-time data collection is performed on traffic, power, and environmental sides; the traffic data includes the traffic network, and the power data includes the power distribution network. Based on the transportation network, the power distribution network, and the virtual links of the charging stations, construct an extended network model of the transportation network; Based on the aforementioned extended network model of the transportation network, a generalized path cost function for heterogeneous users is constructed; the heterogeneous users are fuel vehicle users, travel time-sensitive users, charging cost-sensitive users, and range anxiety-inducing users. Based on the theory of stochastic user equilibrium, the Logit discrete choice model is used to describe the probabilistic joint choice behavior of the heterogeneous users and output the charging flow distribution of each charging station. Based on the traffic-side data, power-side data, environmental-side data, and the charging flow distribution, the optimal power flow model of the distribution network is obtained by combining the second-order cone relaxation algorithm; the optimal power flow model of the distribution network is solved under electrical safety constraints to obtain the operating state of the distribution network; Based on the optimal power flow model of the distribution network, the electrical safety constraints, and the operating status of the distribution network, and combined with the improved alternating direction multiplier method, the guiding price of electricity for charging stations is dynamically generated.

[0006] In one possible implementation, the traffic-side data includes at least regional traffic topology, vehicle origin-destination demand, maximum road network capacity, and free-flow time; the power-side data includes at least distribution network topology parameters, photovoltaic output forecast curves, energy storage capacity and operating boundaries, node voltage and line load margin, maximum charging station capacity, and average charging time; and the environmental-side data includes at least weather forecast data.

[0007] In one possible implementation, constructing an extended network model of the transportation network based on the transportation network, the power distribution network, and the charging station virtual link includes: Based on the traffic network and the power distribution network, construct the actual road links; The extended path formed by the actual road links and the charging station virtual links is used as the extended network model of the transportation network; the actual road links include at least free-flow time and traffic capacity, and the charging station virtual links include at least charging service time, station service capacity, and unit charging price.

[0008] In one possible implementation, the generalized path cost function is a generalized path cost function for fuel vehicle users, a generalized path cost function for users sensitive to travel time, a generalized path cost function for users sensitive to charging costs, and a generalized path cost function for users with range anxiety.

[0009] In one possible implementation, the generalized path cost function for the fuel-powered vehicle user is expressed as: ;in, For the generalized path cost function of fuel vehicle users, This represents the unit time value coefficient for users of gasoline-powered vehicles. This represents the total travel time along the actual route. The generalized path cost function for the time-sensitive user is expressed as: ;in, For time-sensitive users, the generalized path cost function is... The time value coefficient for travel services for users sensitive to travel time. The waiting time value coefficient for users sensitive to travel time. For users who are sensitive to travel time, the charging cost sensitivity coefficient is... This represents the total waiting time along the path. The total charging cost along the route.

[0010] In one possible implementation, the generalized path cost function for the charging cost-sensitive user is expressed as: ;in, For charging cost-sensitive users, the generalized path cost function For charging cost-sensitive users, the time value coefficient of travel service. For charging cost-sensitive users, the waiting time value coefficient The charging cost sensitivity coefficient for charging cost-sensitive users; The generalized path cost function for users with range anxiety is expressed as follows: ;in, For users with range anxiety, the generalized path cost function The time value coefficient for travel services for users with range anxiety. For users with battery anxiety, the waiting time value coefficient For users with range anxiety, the charging cost sensitivity index is... , To assign perceived weights to pre-arrival risk and station congestion risk, , Costs related to pre-arrival risks and station congestion risks.

[0011] In one possible implementation, the method based on stochastic user equilibrium theory, using a Logit discrete choice model to describe the probabilistic joint choice behavior of the heterogeneous users, and outputting the charging flow distribution of each charging station includes: The Logit discrete choice model is used to determine the selection probability of the heterogeneous users; The path traffic is determined based on the selection probability and the total user demand; The charging traffic distribution is determined based on the path traffic and link-path correlation variables.

[0012] In one possible implementation, the power distribution operation status includes the charging load of stations on the transportation side and the station load variables on the distribution side; the step of dynamically generating the guiding price of charging station electricity based on the optimal power flow model of the power distribution network, the electrical safety constraints, and the power distribution network operation status, combined with the improved alternating direction multiplier method, includes: Based on the charging load of the site, the load variables of the site, the optimal power flow model of the distribution network, and the electrical safety constraints, an augmented Lagrangian function is constructed. The augmented Lagrangian function is solved iteratively using the improved alternating direction multiplier method to dynamically generate the guiding price.

[0013] In one possible implementation, the augmented Lagrangian function is expressed as: ;in, The feasible region on the transportation side includes all constraints of the transportation network extended network model. The feasible region on the distribution side includes the optimal power flow model of the distribution network and all its constraints. These are the Lagrange multiplier vectors corresponding to the coupling constraints; For adjustment coefficients of multi-objective functions; For penalty parameters and ; Let the objective function of the transportation system be denoted as . Let be the objective function of the power distribution system.

[0014] The technical solutions provided in this application embodiment have at least the following technical effects or advantages: (1) An extended network model of the transportation network considering the heterogeneity of user perception was constructed. By deeply characterizing the multidimensional behavioral preferences of heterogeneous groups such as those sensitive to travel time, those sensitive to charging costs, and those with range anxiety, it breaks through the limitations of traditional technologies that assume homogeneous groups and cannot accurately capture the chain dynamic effect of "perceptual heterogeneity - path and charging selection - local over-limit of distribution network". Based on the stochastic user equilibrium and Logit discrete choice model, it realizes the accurate characterization of the spatiotemporal evolution law of electric vehicle charging load, and provides a more realistic joint choice behavior modeling tool for the analysis of power-transportation coupled systems. (2) Compared with the traditional centralized optimization method, by introducing the second-order cone relaxation treatment of the optimal power flow of the distribution network, under the premise of strictly ensuring the privacy of local information on the traffic side and the distribution side, the road congestion and local overload of the power grid caused by the excessive concentration of electric vehicles in space are effectively avoided, and the fine guidance and coordinated control capability of complex coupled systems is significantly improved. (3) Based on the improved alternating direction multiplier method, the distributed collaborative solution algorithm solves the bottleneck of high computational complexity and insufficient scalability in traditional large-scale application scenarios by introducing over-relaxation update, adaptive penalty parameter and dynamic electricity price smoothing mechanism; with the aggregated load of each charging station and the dynamic guidance electricity price as the core coupling variables, the algorithm performs double-sided alternating iterative update, providing innovative distributed methodological support for the refined management and intelligent integration of large-scale power-transportation coupling systems. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments of this application or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart of a control method for an electric-transportation coupled system based on the alternating direction multiplier method provided in this application embodiment; Figure 2 A schematic diagram of a vehicle-road-station-network collaborative architecture provided for an embodiment of this application; Figure 3 A schematic diagram of a coupled network of the Nguyen transportation network and the IEEE 33 power distribution system is provided for embodiments of this application. Figure 4 A schematic diagram illustrating a comparison of vehicle traffic flow at a charging station, provided as an embodiment of this application; Figure 5 This application provides a schematic diagram of path traffic for different users under different OD pairs in an embodiment of the present application. Figure 6 A schematic diagram illustrating the convergence process of an improved alternating direction multiplier method provided in this application embodiment; Figure 7 This is a schematic diagram of dynamic electricity pricing for a charging station, provided as an embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0018] In the description of the embodiments of this application, it should be noted that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the embodiments of this application and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application. The terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; a mechanical connection or an electrical connection; a direct connection or an indirect connection through an intermediate medium; or a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in the embodiments of this application according to the specific circumstances.

[0019] Figure 1 This is a flowchart illustrating a control method for an electric-transportation coupled system based on the alternating direction multiplier method, provided as an embodiment of this application. Figure 1 As shown, the method may include the following steps.

[0020] S101. Based on the vehicle-road-station-network collaborative architecture, real-time data collection is performed on traffic-side data, power-side data, and environmental-side data. The traffic-side data includes the traffic network, and the power-side data includes the power distribution network.

[0021] For example, refer to Figure 2 The vehicle-road-station-network collaborative architecture is uniformly controlled by the control center. Among them, "vehicle" can be GVs (Gasoline Vehicles) and EVs (Electric Vehicles); "road" refers to traffic roads; "station" can include SVCs (Static Var Compensators) and charging stations, which are used to stabilize power and provide vehicle charging services, respectively; "network" can include photovoltaic, energy storage and distribution networks, which are used to supply power to SVCs and charging stations.

[0022] S102. Based on the transportation network, the power distribution network, and the virtual links of the charging stations, construct an extended network model of the transportation network.

[0023] S103. Based on the extended network model of the transportation network, construct the generalized path cost function for heterogeneous users; these heterogeneous users include fuel vehicle users, travel time-sensitive users, charging cost-sensitive users, and range anxiety-inducing users.

[0024] S104. Based on the stochastic user equilibrium theory, the Logit discrete choice model is used to describe the probabilistic joint choice behavior of the heterogeneous user and output the charging flow distribution of each charging station.

[0025] S105. Based on the traffic-side data, power-side data, environmental-side data, and the charging flow distribution, the optimal power flow model of the distribution network is obtained by combining the second-order cone relaxation algorithm; the optimal power flow model of the distribution network is solved under electrical safety constraints to obtain the operating state of the distribution network.

[0026] S106. Based on the optimal power flow model of the distribution network, the electrical safety constraints, and the operating status of the distribution network, and combined with the improved alternating direction multiplier method, the guiding price of electricity for charging stations is dynamically generated.

[0027] Through the above technical solutions, an extended network model of the transportation network considering the heterogeneity of user perception was constructed. By deeply characterizing the multidimensional behavioral preferences of heterogeneous groups such as those sensitive to travel time, charging costs, and range anxiety, it overcomes the limitations of traditional technologies that rely on multiple assumptions about homogeneous groups and cannot accurately capture the chain dynamic effect of "perceptual heterogeneity - path and charging selection - local over-limitation of distribution network". Based on stochastic user equilibrium and Logit discrete choice models, it achieves a precise characterization of the spatiotemporal evolution of electric vehicle charging load, providing a more realistic joint choice behavior modeling tool for the analysis of power-transportation coupled systems.

[0028] In one possible implementation, the traffic-side data includes at least regional traffic topology, vehicle origin-destination demand, maximum road network capacity, and free-flow time; the power-side data includes at least distribution network topology parameters, photovoltaic output forecast curves, energy storage capacity and operating boundaries, node voltage and line load margin, maximum charging station capacity, and average charging time; and the environmental-side data includes at least weather forecast data.

[0029] For example, the traffic topology of the area can be obtained through map software, and network characteristics such as vehicle origin-destination (OD) demand, maximum road network capacity, and free-flow time can be obtained through vehicle terminals and roadside units; in addition, weather forecast data can be collected through sensors.

[0030] In one possible implementation, S102 may include: constructing actual road links based on the transportation network and the power distribution network; using the extended path formed by the actual road links and the charging station virtual links as the extended network model of the transportation network; the actual road links include at least free-flow time and traffic capacity, and the charging station virtual links include at least charging service time, station service capacity, and unit charging price.

[0031] For example, to provide a unified description of road travel and charging services, charging stations are abstracted as virtual road segments embedded in the transportation network, and charging at a station is considered a special passage in the electric vehicle's path, thus establishing a virtual link for charging stations. Therefore, the travel path of an electric vehicle from its origin to its destination is no longer just a combination of several actual road links, but an extended path composed of both actual road links and virtual links for charging stations.

[0032] For example, the extended transportation network link set can be defined as follows: To reflect the charging behavior of electric vehicles, it is further divided into a set of actual road links and a set of virtual charging station links, namely: ;in, It represents the actual set of road links, describing the actual operation of vehicles in the traffic network; This represents a set of virtual links for charging stations, describing the service process of electric vehicles completing charging within the charging station.

[0033] set up Represents any actual road link. This represents any virtual link at a charging station. For a real road link, its main attributes include free-flow time and capacity; for a virtual link at a charging station, its main attributes include charging service time, station service capacity, and unit charging price. Using this extended network representation, the time loss, queuing loss, and charging cost of electric vehicles on the path can be uniformly mapped to the link impedance.

[0034] Reference Figure 3 The power distribution system diagram is shown in the coupled network diagram. For the transportation side, the Nguyen road network is used as the transportation network, which includes 13 nodes and 19 actual road links. This transportation network and the power distribution network are spatially coupled through 4 charging stations. The transportation location and the power distribution access location are shown in the diagram. Figure 3 As shown, to describe the charging behavior of electric vehicles during their journey, four virtual charging station links (CS1, CS2, CS3, and CS4) are added to the existing transportation network, forming an extended transportation network that includes charging behavior. Regarding traffic demand settings, the scenario includes four vehicle origin-destination demand pairs (OD pairs): 1→2, 1→3, 4→2, and 4→3, with each OD pair having a travel demand of 100veh / h. The vehicles consist of gasoline-powered vehicles and three types of electric vehicles, with gasoline-powered vehicles accounting for 70% of the total demand, and time-sensitive electric vehicles, charging cost-sensitive electric vehicles, and range-anxiety-inducing electric vehicles each accounting for 10%.

[0035] In one possible implementation, the generalized path cost function is a generalized path cost function for fuel vehicle users, a generalized path cost function for users sensitive to travel time, a generalized path cost function for users sensitive to charging costs, and a generalized path cost function for users with range anxiety.

[0036] For example, considering the significant heterogeneity in the perceptions of electric vehicle users, they can be divided into three categories: users sensitive to travel time, users sensitive to charging costs, and users with range anxiety. These users will be included in a unified allocation system along with traditional gasoline vehicle users. Gasoline vehicle users will only compete for road resources, not charging resources; the three categories of electric vehicle users will compete for both road resources and charging service resources.

[0037] For the sake of consistency, the user category set is defined as follows: ;in, This refers to users of gasoline-powered vehicles; This indicates users who are sensitive to travel time; This indicates users who are sensitive to charging costs; This indicates users with battery life anxiety.

[0038] like Figure 4 As shown in the figure, the path distribution traffic of traditional fuel vehicles and three types of electric vehicles (EVs) with different preferences at each charging station is illustrated from the starting point to the destination. The results show significant differences in path selection patterns among different users. The traffic flow of traditional fuel vehicles is relatively dispersed across all paths, while electric vehicles exhibit a high concentration on a few advantageous paths. The charging traffic distribution at the four charging stations shows significant differences. Overall, CS2 and CS3 handle the majority of charging demand, accounting for approximately 56% of the total charging traffic combined, indicating a clear clustering characteristic in the charging load under equilibrium conditions. By type, CS3 has the highest proportion of cost-sensitive (price-sensitive) users (approximately 48%), making it the most attractive station for EVCs. CS4 has approximately 45% of range-anxiety users, making it the dominant type at the station, and its spatial location and congestion level better align with EVAs' preference for safe charging. CS2 has a relatively even proportion of the three types of charging, indicating that it is attractive to different types of users. In contrast, CS1 has the lowest total traffic, accounting for only about 19% of the total charging traffic, making its overall service attractiveness relatively weak. This demonstrates that the generalized path cost function described above can accurately capture and quantify the diverse decision-making behaviors of users of various types of electric vehicles, thus providing a reliable user behavior model foundation for the refined guidance of the coupled system.

[0039] In some embodiments, an impedance model of the actual road link and the virtual link of the charging station can be established first, and then a generalized path cost function can be constructed based on the impedance model; the construction process of the impedance model may include the following steps.

[0040] For any real road link The classic BPR (Bureau of Public Roads) function is used to describe the link travel time under road congestion conditions, i.e.: ;in, , These represent link impedance parameters, reflecting the intensity of congestion impact and the degree of nonlinear growth, respectively. Indicates free-flow time; Indicates link capacity; This represents the total link traffic, generated by the competition for road resources among four types of users. When link traffic is low... The actual travel time of the link is relatively small, and it is close to the free-flow time. However, when the link traffic gradually approaches or exceeds the link capacity, the congestion effect is significantly enhanced, and the link travel time will increase non-linearly.

[0041] Total dwell time on the virtual link of the charging station It can be represented as: ;in, This represents the average service time required to complete one charge at station s; Indicates electric vehicle users at the station s Average queue waiting time; This indicates the total charging flow reaching this station.

[0042] Let the set of OD pairs be denoted as For any OD pair Its total travel demand is denoted as After considering the differences in user types, the OD requirements are divided into four parts: ;in: —Demands of gasoline vehicle users; — Needs of users sensitive to travel time; —The needs of charging cost-sensitive users; —Needs of users with range anxiety. The above formula shows that travelers under the same OD pair are not a homogeneous group, but rather a combination of users with various behavioral preferences; the behavioral characteristics of gasoline vehicle users are relatively simple, and their route selection is mainly affected by road travel time; users sensitive to travel time pay more attention to the overall travel efficiency and waiting time at stations; users sensitive to charging costs tend to choose stations with lower prices, even if it requires some detours and waiting; users with range anxiety, in addition to paying attention to time and cost, also show greater sensitivity to the distance to the destination and the risk of congestion at the target station.

[0043] In one possible implementation, the generalized path cost function for the fuel-powered vehicle user is expressed as: ;in, For the generalized path cost function of fuel vehicle users, This represents the unit time value coefficient for users of gasoline-powered vehicles. The total travel time on the actual path; the generalized path cost function for this travel time-sensitive user is expressed as: ;in, For time-sensitive users, the generalized path cost function is... The time value coefficient for travel services for users sensitive to travel time. The waiting time value coefficient for users sensitive to travel time. For users who are sensitive to travel time, the charging cost sensitivity coefficient is... This represents the total waiting time along the path. The total charging cost along the route.

[0044] In one possible implementation, the generalized path cost function for the charging cost-sensitive user is expressed as: ;in, For charging cost-sensitive users, the generalized path cost function For charging cost-sensitive users, the time value coefficient of travel service. For charging cost-sensitive users, the waiting time value coefficient The charging cost sensitivity coefficient is used for charging cost-sensitive users; the generalized path cost function for this range-anxiety user is expressed as: ;in, For users with range anxiety, the generalized path cost function The time value coefficient for travel services for users with range anxiety. For users with battery anxiety, the waiting time value coefficient For users with range anxiety, the charging cost sensitivity index is... , To assign perceived weights to pre-arrival risk and station congestion risk, , Costs related to pre-arrival risks and station congestion risks.

[0045] For example, for users of gasoline-powered vehicles, their travel does not involve charging; therefore, their generalized path cost is determined solely by road travel time, and their behavioral logic aligns with traditional traffic allocation models. For users sensitive to travel time, they are more concerned with overall travel efficiency and waiting losses. Generally, these users are more willing to pay higher economic costs to shorten travel time and reduce queuing, thus exhibiting higher [cost / effort]. and ,and Relatively low. For users sensitive to charging costs, who are more concerned about the cost of recharging, for this type of user, Often significantly higher than and This indicates that they are more sensitive to price differences at charging stations and are more likely to choose charging stations with lower electricity prices, even if it requires some time to detour or wait. For users with range anxiety, in addition to the costs of travel, waiting, and fees, they also need to consider the risks of charging before reaching a station and the uncertainty of high-load charging stations.

[0046] In one possible implementation, S104 may include: using the Logit discrete choice model to determine the selection probability of the heterogeneous user; determining the path traffic based on the selection probability and the total user demand; and determining the charging traffic distribution based on the path traffic and link-path correlation variables.

[0047] For example, the above process may include the following steps.

[0048] (1) The Logit discrete choice model based on stochastic user equilibrium is used to describe the random choice behavior of various users among candidate paths. For OD pairs w The following categories m User, path The selection probability is defined as: ;in, Indicates category m User selects path p The probability, Indicates category m User selects path p Generalized path cost Indicates category m This is a cost-sensitive parameter for users. The formula indicates that the lower the path cost, the greater the probability of it being chosen by users; however, due to perceptual errors and behavioral randomness, even if it's not the lowest-cost path, it may still be chosen by some users, thus making the path flow distribution exhibit discrete probability characteristics. (Parameter) The size reflects the user's sensitivity to cost differences. The larger the scale, the more users tend to focus on low-cost paths; The smaller the value, the more dispersed the user choices are.

[0049] (2) OD to w Next category m User in path p The path flow can be represented as the product of total demand and the probability of choice: ;in, Indicates category m User selects path p Path traffic.

[0050] (3) To achieve the conversion from path flow to link flow, define link-path association variables: Therefore, the category m User in the link a The charging flow distribution on the surface can be represented as: .

[0051] like Figure 5 As shown in the figure, the path flow distribution of traditional gasoline vehicles (GV) and three types of electric vehicles (EV, including time-sensitive EVT, charging cost-sensitive EVC, and range-anxiety-sensitive EVA) on four different origin-destination pairs (OD pairs) is intuitively displayed. The distribution results show that the flow distribution of traditional gasoline vehicles on each candidate path is relatively smooth and dispersed, as their choices are only affected by road congestion. The flow of electric vehicles is highly concentrated on a few advantageous paths. Furthermore, on the same path, the proportions of time-sensitive (EVT), charging cost-sensitive (EVC), and range-anxiety-sensitive (EVA) users are significantly different. Thus, by using the Logit discrete choice model, the psychological cost formula for heterogeneous users can be transformed into actual traffic behavior and path flow, accurately and reliably outputting the spatiotemporal evolution patterns of multiple vehicle types in the coupled network.

[0052] In another possible implementation, S105 may include: establishing an optimal power flow model of the distribution network based on the branch power flow model, and performing second-order cone relaxation processing on the model.

[0053] (1) The application site charging demand mapping model maps the charging station flow to the distribution network node load. If the charging station s Access to distribution network nodes i Based on the traffic-side path flow results, the study period Internal site s The total charging power requirement is: ;in, Indicates a site-path related variable; Indicates category m User along the pathp On the site s Average charging capacity; Indicates category m User selects path p Path traffic.

[0054] Site s Average charging active power during the study period It can be represented as: If the charging equipment operates at a constant power factor, then the station s Its reactive power This can be further expressed as: ;in, Represented as a site s The power factor angle. Therefore, the node i The total charging load formed by the connection of the charging station is: , .

[0055] (2) Based on the above, establish an optimal power flow operation model for the distribution network after second-order cone relaxation: Using directed graphs To describe the distribution network topology, where, E N and E L Let these represent the set of nodes and the set of lines in the distribution network, respectively. Given the network topology... G E and electrical parameters of the line At that time, a second-order cone-relaxed DistFlow power flow model is used to describe the operational constraints of the PDS: ; ; ; ; ; ; ; in, express t Constant flow into the line Active and reactive power at the head end; and For the line Resistance and reactance; and express t Flowing nodes at all times j Active power and reactive power,k It is a node j Downstream nodes; and It is the node injected power; , This indicates the active and reactive power injected into the distribution network from the upper-level power grid. Indicates access node j Photovoltaic injection power; , Indicates access node j The energy storage discharge power and charging power; and Indicates access node j The normal active and reactive loads; Indicates access node j The charging load; Indicates access node j The charging load; Represents a node i To the node j Outflowing current; and respectively t Time Node j and nodes i The voltage amplitude.

[0056] (3) Considering the safe and stable operation of the power grid and the requirements for power quality, an electrical safety constraint is set here, which can be expressed as: , ;in, , These are the minimum and maximum allowable values ​​for the node voltage during operation, respectively. This represents the maximum allowable line current during operation.

[0057] In one possible implementation, the distribution operation state includes the charging load of stations on the traffic side and the station load variables on the distribution side; S106 may include: constructing an augmented Lagrangian function based on the station charging load, the station load variables, the optimal power flow model of the distribution network, and the electrical safety constraints; iteratively solving the augmented Lagrangian function using the improved alternating direction multiplier method to dynamically generate the pilot price. For example, the distribution operation state may also include node voltage states, branch power distribution, network losses, and generation / load matching, etc., which are not limited here.

[0058] In one possible implementation, the augmented Lagrangian function is expressed as: ;in, The feasible region on the transportation side includes all constraints of the transportation network extended network model. The feasible region on the distribution side includes the optimal power flow model of the distribution network and all its constraints. These are the Lagrange multiplier vectors corresponding to the coupling constraints; For adjustment coefficients of multi-objective functions; For penalty parameters and ; Let the objective function of the transportation system be denoted as . Let be the objective function of the power distribution system. The constraints of the optimal power flow model for this power distribution network can include node power balance, node voltage constraints, branch current constraints, branch capacity constraints, and second-order cone constraints, etc., which are not limited here.

[0059] For example, the charging load at a transportation-side station can be represented as The site load variable used on the distribution side can be expressed as: The load of each site is organized into a vector form: ; .

[0060] To ensure bilateral cognitive consistency, the following augmented Lagrangian function can be constructed: .

[0061] The basic iterative structure of the standard improved alternating direction multiplier method can be expressed as: ; ; .

[0062] For example, you can refer to Figure 6 As shown in (a), the improved Alternating Direction Method of Multipliers (ADMM) can simultaneously satisfy the thresholds for both the original residual and the dual residual after approximately 21 iterations, while the standard Alternating Direction Method... Figure 6 (b) requires approximately 99 iterations to reach convergence, while the improved strategy significantly reduces the number of coordination iterations. This is because the method in this application does not have a fixed penalty parameter, but rather adaptively adjusts the penalty parameter based on the relative magnitudes of the original and dual residuals. This achieves a dynamic balance between accelerating feasibility approximation and suppressing variable oscillations. Especially in the initial coupling phase, there is a strong inconsistency between the load response of the charging price on the traffic side and the power supply boundary on the distribution side. If the penalty parameter... If the penalty parameter is too small, the original residual will decrease slowly; if the penalty parameter is too small... If the value is too large, it can easily cause excessive fluctuations in the dual variable.

[0063] In some embodiments, the solution process of the improved alternating direction multiplier method described above may include the following steps.

[0064] (1) Initialization. First, the initial combined traffic flow is given. Site load And the initial feasible solution on the distribution side that satisfies the constraints of the optimal power flow model for the distribution network. Thus, the initial site load is obtained. Simultaneously initialize the dual variables of the improved alternating direction multiplier method. Penalty parameters Over-relaxation coefficient And given the original residual threshold Dual residual threshold and the threshold for relative change in the objective function .

[0065] (2) Solving the traffic subproblem. Given... Then, the traffic subproblem is solved iteratively using the inner near-end MSA-Logit algorithm. The specific steps of the MSA-Logit algorithm are as follows.

[0066] Step a: Calculate the segment flow based on the current path flow. And use the BPR function to update the road segment time. This allows us to obtain the travel time for the path-site combination. .

[0067] Step b: Based on the current station arrival volume Update the waiting time of each site using the Davidson function. And the range anxiety cost is updated based on the distance to the vehicle's target station.

[0068] Step c: Calculate the equivalent generalized cost of each path-site combination. .

[0069] Step d: Update the joint choice probability using the Logit discrete choice model: And calculate the auxiliary flow: .

[0070] Step e: Update traffic flow using MSA: ;in, Let be the number of iterations within the traffic subproblem.

[0071] Step f: If the relative error of the inner traffic layer satisfies: If the inner traffic iteration stops, output the traffic side solution. and site load Otherwise, return to step a.

[0072] (3) Solving the electronic distribution problem. Update the results on the given traffic side. The problem of distributing electrons is then written as: ; Its constraints are SOCP (Second-Order Cone Programming) power flow constraints, namely node power balance, node voltage constraints, branch current constraints, branch capacity constraints, and second-order cone constraints. After solving the power distribution problem, we obtain: .

[0073] (4) Over-relaxation update. To reduce the oscillation of coupled variables during the alternating update process on both sides, an over-relaxation mechanism is introduced after the load update on the distribution side. The over-relaxed station load is defined as: In the formula, The relaxation coefficient is represented, and .when At times, the consensus variables can be accelerated towards the new solution; if the residuals increase, the consensus variables can also be adjusted accordingly. Adjust it back to a value close to 1.

[0074] (5) Dual variable update. Based on the over-relaxed distribution-side load, update the dual variables: .

[0075] (6) Dynamic electricity guidance price Update. Charging station. s The base electricity price is Set the price mapping coefficient to Then the first k The original dynamic electricity price generated by the multiplier vector in the next outer iteration can be expressed as: ;in, Indicates site s In the k The corresponding improved alternating direction multiplier method multiplier under the second outer layer iteration; This indicates that the price range is projected based on the displayed electricity price boundary constraints. The limiting operator.

[0076] Let the smoothing coefficient be... Then the site s In the k The actual electricity price under the next outermost iteration is: When a power supply to a site is tight in the previous iteration, the corresponding multiplier will be adjusted to increase the electricity price at that site, thereby discouraging users from continuing to concentrate at that site. Conversely, when a power supply to a site is abundant, the pressure on electricity prices will be reduced, which will help attract more users to switch to that site for energy replenishment.

[0077] (7) Adaptive penalty parameter update. Define the first... k+ The original residual and dual residual of the first iteration are as follows: , .

[0078] The penalty parameters are updated as follows: In the formula, , This represents the increase or decrease coefficient of the penalty parameter, and , ; This represents the residual balance discriminant coefficient, and This strategy can strengthen consistency constraints when the original residuals are large, and reduce the rigidity of the penalty term when the dual residuals are large, thereby improving the overall convergence efficiency.

[0079] like Figure 7 As shown, this reflects the coordinated process of price signals from each charging station guiding the redistribution of traffic charging demand, and the distribution side responding to supply according to network constraints. The dynamic electricity price is determined by the established electricity price update formula and is subject to boundary constraints of 0.5-2.0 yuan / kWh. Figure 7 The electricity prices at various charging stations in the region all exhibit a smooth upward trend and gradually stabilize, without significant fluctuations. This demonstrates the effectiveness of the electricity price smoothing and adaptive penalty parameters used in this application. The setting effectively suppressed the drastic fluctuations in the iteration of the improved alternating direction multiplier method.

[0080] (8) Outer stopping criterion. The algorithm is considered convergent when both of the following conditions are met: , ;in, , Represented as the original residual and the threshold of the dual residual.

[0081] The various embodiments in this specification are described in a progressive manner. For the same or similar parts between the various embodiments, please refer to each other. Each embodiment focuses on describing the differences from other embodiments.

[0082] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of this application.

Claims

1. A control method for an electric-transportation coupled system based on the alternating direction multiplier method, characterized in that, The method includes: Based on a vehicle-road-station-network collaborative architecture, real-time data collection is performed on traffic, power, and environmental sides; the traffic data includes the traffic network, and the power data includes the power distribution network. Based on the transportation network, the power distribution network, and the virtual links of the charging stations, an extended network model of the transportation network is constructed. Based on the aforementioned extended network model of the transportation network, a generalized path cost function for heterogeneous users is constructed; the heterogeneous users are fuel vehicle users, travel time-sensitive users, charging cost-sensitive users, and range anxiety-inducing users. Based on the theory of stochastic user equilibrium, the Logit discrete choice model is used to describe the probabilistic joint choice behavior of the heterogeneous users and output the charging flow distribution of each charging station. Based on the traffic-side data, power-side data, environmental-side data, and the charging flow distribution, the optimal power flow model of the distribution network is obtained by combining the second-order cone relaxation algorithm; the optimal power flow model of the distribution network is solved under electrical safety constraints to obtain the operating state of the distribution network; Based on the optimal power flow model of the distribution network, the electrical safety constraints, and the operating status of the distribution network, and combined with the improved alternating direction multiplier method, the guiding price of electricity for charging stations is dynamically generated.

2. The method according to claim 1, characterized in that, The traffic-side data includes at least the regional traffic topology, vehicle origin-destination demand, maximum road network capacity, and free-flow time; the power-side data includes at least the distribution network topology parameters, photovoltaic output forecast curves, energy storage capacity and operating boundaries, node voltage and line load margin, maximum charging station capacity, and average charging time; and the environmental-side data includes at least weather forecast data.

3. The method according to claim 1, characterized in that, The step of constructing an extended network model of the transportation network based on the transportation network, the power distribution network, and the virtual links of charging stations includes: Based on the traffic network and the power distribution network, construct the actual road links; The extended path formed by the actual road links and the charging station virtual links is used as the extended network model of the transportation network; the actual road links include at least free-flow time and traffic capacity, and the charging station virtual links include at least charging service time, station service capacity, and unit charging price.

4. The method according to claim 1, characterized in that, The generalized path cost function is defined as the generalized path cost function for fuel vehicle users, the generalized path cost function for users sensitive to travel time, the generalized path cost function for users sensitive to charging costs, and the generalized path cost function for users with range anxiety.

5. The method according to claim 4, characterized in that, The generalized path cost function for fuel-powered vehicle users is expressed as follows: ;in, For the generalized path cost function of fuel vehicle users, This represents the unit time value coefficient for users of gasoline-powered vehicles. This represents the total travel time along the actual route. The generalized path cost function for the time-sensitive user is expressed as: ;in, For time-sensitive users, the generalized path cost function is... The time value coefficient for travel services for users sensitive to travel time. The waiting time value coefficient for users sensitive to travel time. For users who are sensitive to travel time, the charging cost sensitivity coefficient is... This represents the total waiting time along the path. The total charging cost along the route.

6. The method according to claim 5, characterized in that, The generalized path cost function for the charging cost-sensitive user is expressed as follows: ;in, For charging cost-sensitive users, the generalized path cost function For charging cost-sensitive users, the time value coefficient of travel service. For charging cost-sensitive users, the waiting time value coefficient The charging cost sensitivity coefficient for charging cost-sensitive users; The generalized path cost function for users with range anxiety is expressed as follows: ;in, For users with range anxiety, the generalized path cost function The time value coefficient for travel services for users with range anxiety. For users with battery anxiety, the waiting time value coefficient For users with range anxiety, the charging cost sensitivity index is... , To assign perceived weights to pre-arrival risk and station congestion risk, , Costs related to pre-arrival risks and station congestion risks.

7. The method according to claim 1, characterized in that, The method, based on stochastic user equilibrium theory, uses a Logit discrete choice model to describe the probabilistic joint choice behavior of heterogeneous users, and outputs the charging flow distribution of each charging station, including: The Logit discrete choice model is used to determine the selection probability of the heterogeneous users; The path traffic is determined based on the selection probability and the total user demand; The charging traffic distribution is determined based on the path traffic and link-path correlation variables.

8. The method according to claim 1, characterized in that, The power distribution operation status includes the charging load of stations on the transportation side and the station load variables on the distribution side; the step of dynamically generating the guiding price of charging station electricity based on the optimal power flow model of the power distribution network, the electrical safety constraints, and the power distribution network operation status, combined with the improved alternating direction multiplier method, includes: Based on the charging load of the site, the load variables of the site, the optimal power flow model of the distribution network, and the electrical safety constraints, an augmented Lagrangian function is constructed. The augmented Lagrangian function is solved iteratively using the improved alternating direction multiplier method to dynamically generate the guiding price.

9. The method according to claim 8, characterized in that, The augmented Lagrange function is expressed as: ;in, The feasible region on the transportation side includes all constraints of the transportation network extended network model. The feasible region on the distribution side includes the optimal power flow model of the distribution network and all its constraints. These are the Lagrange multiplier vectors corresponding to the coupling constraints; For adjustment coefficients of multi-objective functions; For penalty parameters and ; Let the objective function of the transportation system be denoted as . Let be the objective function of the power distribution system.

Images