Charging facility economic optimization method based on electric vehicle charging demand management

Abstract

The invention discloses a charging facility economic optimization method based on electric vehicle charging demand management, and the method comprises the following steps: building a transportation-distribution network charging distribution model based on the number of charging facilities and charging vehicles of a charging station in a certain region, determining the optimal positions of the charging facilities in a traffic system and an electric power system, and carrying out the charging demand management of the charging facilities. Estimating the capacity and the service rate of the charging station; based on a transportation-distribution network charging distribution model, designing an incentive method for changing a charging behavior, and reducing the utilization rate of peak periods and high-demand places; constructing a double-layer optimization problem, and carrying out demand distribution on a user among a plurality of charging stations so as to reduce waiting time; according to the method, a double-layer optimization problem is converted into a single-stage optimization problem, a target function of a lower-layer problem is relaxed, the double-layer optimization problem is converted into a non-convex MINLP optimization problem, the non-convex MINLP optimization problem is solved, and maximum utilization of charging requirements of all charging stations in a certain area is achieved.

Description

Technical Field

[0001] The present invention belongs to the cross - field of transportation engineering and electrical engineering, and relates to an economic optimization method for charging facilities based on electric vehicle charging demand management. Background Art

[0002] Recent trends indicate that the electric vehicle (EV) market has experienced significant growth in the past decade and is expected to continue growing in the coming years. Therefore, policymakers are paying more attention to expanding charging infrastructure to support the increasing number of electric vehicles on the road. However, the availability of fast chargers is still limited, and adding more chargers may double the peak power demand. To ensure the quality and stability of infrastructure services, it is crucial to implement a comprehensive demand management plan. With the rapid growth of the electric vehicle scale, the differentiated service demands of users have become increasingly prominent, and the charging operation market situation has also changed greatly. This is mainly manifested in: the profitability of the operating vehicle market is gradually declining, the market development tends to be saturated, and the charging operation mode is relatively stable; the private passenger vehicle charging market is gradually segmented according to driving habits and charging behaviors, and specialized operating companies have emerged in different segmented fields.

[0003] Problems existing in the current construction of electric vehicle charging infrastructure:

[0004] (1) First, in terms of the coordination of planning and layout, there is a structural contradiction between the actual number and utilization rate of charging piles being built. On the one hand, car owners complain about the lack of charging piles, and on the other hand, operators are worried about the low utilization rate of charging piles. This is because there is a mismatch in space and structure between the planning and construction of charging piles and the use of new energy vehicles.

[0005] (2) From the perspective of charging implementation operators, the business model needs to be further explored. Currently, operators only make a profit by charging a charging service fee (charged per kilowatt - hour), which is difficult to reflect the concept of charging services. The core value of a charging pile lies in its location, and there are very large differences in construction costs such as land prices. However, currently, whether it is charging in the urban core area or in the suburbs, the service fee charged according to the electricity consumption is the same.

[0006] (3) From the perspective of car owners, what they complain about more is the poor charging experience: one is the difficulty in finding a charging pile, the second is the long charging time, and there are also cumbersome procedures for registering different APPs. The charging anxiety still exists.

[0007] The current charging plan for electric vehicles needs to make an overall plan for the construction of charging piles based on factors such as urban development plans, traffic flow distribution, and power supply conditions. Ensure the reasonable layout of charging piles in areas such as residential areas, commercial areas, public parking lots, and highway service areas to meet the charging needs of electric vehicle users. Specify the technical specifications, power levels, interface standards, etc. of charging piles. For example, unify the charging interface standard to ensure that electric vehicles of different brands and models can use the charging piles for charging. At the same time, corresponding standards have also been formulated for the safety performance, protection level, etc. of charging piles to ensure the safety and reliability of the charging process. Summary of the Invention

[0008] To solve the above problems, the technical solution adopted by the present invention is: a charging facility economic optimization method based on electric vehicle charging demand management, including the following steps:

[0009] Based on the number of charging facilities and charging vehicles in the charging stations in a certain area, establish a transportation-allocation network toll allocation model to determine the optimal locations of charging facilities in the transportation system and the power system, and estimate the capacity and service rate of the charging stations;

[0010] Based on the transportation-allocation network toll allocation model, design an incentive method to change charging behavior and reduce the utilization rate during peak hours and at high-demand locations;

[0011] Construct a two-layer optimization problem to allocate the demand among multiple charging stations by users to reduce the waiting time;

[0012] Convert the two-layer optimization problem into a single-level optimization problem. By relaxing the objective function of the lower-layer problem, convert the two-layer optimization problem into a non-convex MINLP optimization problem, and solve the non-convex MINLP optimization problem to maximize the utilization of the charging demand of each charging station in a certain area.

[0013] Furthermore: The transportation-allocation network toll allocation model includes a traffic model and a distribution network model;

[0014] The expression of the traffic model is as follows:

[0015]

[0016] Where: represents the maximum distance between two consecutive stations, N f represents a set of nodes in the power grid that can cover path f and are sorted by distance. Since the driving range of electric vehicles is limited, the dispersion of charging locations also needs to be considered. r fis a binary variable. If the path is f, its value is 1. Equation (1) ensures that the maximum distance between two consecutive stations does not exceed the driving range L. Equation (2) guarantees that the interval between stations does not exceed L. Equation (3) requires that at least one station is needed near the origin within L / 2 to meet the route requirements;

[0017] The expression of the distribution network model is as follows:

[0018]

[0019] Where: represents the actual power injected by node h in the power grid, and α t represents the hourly electricity price at time t, represents the node power consumption, which is obtained from the basic load and equivalent power load during the charging period. The power load of each charging station is obtained from the number of charging vehicles, SOC, charging rate, battery capacity, and travel distance within each time period. and are the reactive power and injected power load of each node respectively. Equation (6) controls the node voltage of the power grid according to the distribution line data. The power grid meets the demand by providing consistent power, thereby regulating the performance of the charging station and realizing the management of steady-state and continuous current for the power grid of each power station.

[0020] Furthermore: The incentive method for changing charging behavior predicts the driver's reaction to changes in charging time and location. The value of charging incentives affects users' charging behavior by increasing the user's utility level. The formula expression of the utility level is as follows:

[0021]

[0022] The acceptance of the utility of changing the charging behavior is represented by the usage level, that is, the user accepts the incentive for the alternative solution. ψ0 represents the benefit level of habitual charging behavior, is the additional utility change due to switching to an alternative charging station. Under this incentive method, the benefit of changing charging behavior is greater than the benefit under habitual behavior.

[0023] Furthermore: The bilevel optimization problem includes an upper-level problem and a lower-level problem. The upper-level problem is to improve the utilization rate of the station by selecting the best incentive plan,

[0024] The expression of the upper-level problem is as follows:

[0025]

[0026] Where, is the service rate of charger k at time t, Denote the number of chargers \(k\) in charging station \(h\), and use Equation (8) to calculate the number of charging times that can be served at station \(t\).

[0027] The lower-level problem is the interaction of each user's choice on the electric vehicle charging demand management system and other users' decisions.

[0028]

[0029] Among them, the charging cost of the user according to the SOC and the charging level is \(C\) i , \(y\) i,h,t is a binary variable, representing the choice of charging user \(i\) for charging station \(h\) and charging time \(t\), and \(x\) h represents the discount level provided by the station, and Equation (10) calculates the minimum charging cost of the user.

[0030] Furthermore: The process of transforming the bilevel optimization problem into a single-level optimization problem, by relaxing the objective function of the lower-level problem, transforming the bilevel optimization problem into a non-convex MINLP optimization problem, and solving the non-convex MINLP optimization problem to maximize the utilization of the charging demands of each charging station in a certain area is as follows:

[0031] First, in the first stage, taking it as the only variable, use a heuristic algorithm to solve; here, a genetic algorithm is adopted, and the specific process is as follows:

[0032] First, perform variable selection, only consider the incentive level \(x\) as the variable, and operate on it as the chromosome in the genetic algorithm;

[0033] Then set the fitness function, aiming to minimize the maximum demand allocation rate of the system, and perform iterations. Based on the principle of natural selection, the genetic algorithm approximates the optimal solution by continuously evolving the solutions in the population. In each iteration, by changing the value of the chromosome, that is, the incentive level \(x\), observe the behavioral changes of users' choices based on the charging station capacity and their own preferences. According to the users' choices, calculate the utilization rate of the system, evaluate the fitness of each candidate solution (chromosome), and the solutions with higher fitness are more likely to be retained in the next generation. After multiple iterations, find a set of values of the optimal discount incentive level \(x\) that can encourage users to change their charging habits:

[0034]

[0035] Equation (10) reflects that this process is an iterative process, using the objective function as the survival method. Through the calculations in this stage, the optimal discount level that can encourage users to change their charging habits at each station can be obtained;

[0036] In the second stage, the optimal incentive discounts obtained in the first stage are used to reallocate users to each charging station to find the optimal demand distribution. In this process, factors such as users' travel routes, charging demands, time flexibility, as well as the capacity and incentives of charging stations are considered. The charging choices of users are adjusted through an optimization algorithm.

[0037] Finally, the branch-and-price-and-cut method is adopted to solve the MINLP problem.

[0038] Furthermore, the process of using the branch-and-price-and-cut method to solve the MINLP problem is as follows:

[0039] After determining the user allocation, the problem is transformed into a mixed-integer nonlinear programming (MINLP) problem for solution. In this process, the branch-and-price-and-cut algorithm is used to handle the non-convexity and integer variable constraints of the problem. The branching operation divides the solution space of the problem into multiple branches, searches each subspace, and gradually approaches the optimal solution. The price reduction operation tightens the relaxed solution of the problem and reduces the search space by identifying and exploiting the special structure of the problem, such as valid inequality constraints (cutting planes). The pricing operation involves calculating the lower bound of the optimal solution for each subproblem during the branching process and selecting the most promising branch for further exploration based on the lower bound. By continuously repeating these operations, the optimal solution that satisfies the constraint conditions is finally found, that is, the optimal incentive level x for each charging station at each time period, the charging choices y of users, and the minimum-maximum demand allocation rate z of the system are determined.

[0040] A charging facility economic optimization method based on electric vehicle charging demand management provided by the present invention is an incentive method for electric vehicle charging demand management, aiming at all electric vehicle users, and aims to maximize the utilization rate of the system by effectively allocating the charging demands of each charging station. Firstly, the transportation-allocation network toll allocation model is utilized, considering the number and location of charging facilities. Secondly, appropriate incentives are determined at each station, and barrier-free services are provided for disabled and elderly users to minimize the total waiting time. Finally, the problem is formulated as a bilevel optimization model and transformed into a single-level mixed-integer nonlinear programming, and a method combining a heuristic algorithm and a branch-and-cut algorithm is proposed. The bilevel optimization model proposed by the present invention provides an effective method for the future design of maximizing the use efficiency of electric vehicle charging stations and has the following advantages:

[0041] 1) The present invention optimizes the facility utilization rate, reasonably allocates the charging demand based on traffic and grid load, effectively utilizes the charging facilities, avoids overcrowding at some stations, and improves the overall facility utilization rate.

[0042] 2) The present invention not only balances demand and reduces waiting time, but also improves the social fairness and inclusiveness of charging facilities. By formulating incentive measures based on individual charging behaviors, including charging time and station selection incentives, it balances the charging demand at each station and reduces the user waiting time.

[0043] 3) By establishing a utility behavior model to simulate the response of users to incentives, considering the cost optimization behavior of users and their expectations for the charging service level, it improves user satisfaction and convenience.

[0044] 4) The present invention adopts a method combining a two-layer optimization model with a heuristic algorithm and a branch-and-cut algorithm, which can flexibly adapt to different charging scenarios and user needs, and effectively solve complex optimization problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0046] Figure 1 It is a layout diagram of the charging system management framework of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0047] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine the embodiments to detail the present invention.

[0048] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all of them. The following description of at least one exemplary embodiment is merely illustrative and in no way restrictive of the present invention and its application or use. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0049] Figure 1 It is a layout diagram of the charging system management framework of the present invention.

[0050] A charging facility economic optimization method based on electric vehicle charging demand management includes the following steps:

[0051] S1: Based on the number of charging facilities and charging vehicles at charging stations in a certain area, establish a transportation-allocation network toll allocation model to determine the optimal locations of charging facilities in the transportation system and the power system, and estimate the capacity and service rate of charging stations;

[0052] S2: Based on the transportation-allocation network toll allocation model, design an incentive method to change charging behavior and reduce the utilization rate during peak hours and at high-demand locations;

[0053] S3: Construct a two-layer optimization problem to allocate the demand among multiple charging stations for users to reduce the waiting time;

[0054] S4: Transform the two-layer optimization problem into a single-level optimization problem. By relaxing the objective function of the lower-layer problem, convert the two-layer optimization problem into a non-convex MINLP optimization problem, and solve the non-convex MINLP optimization problem to maximize the utilization of the charging demand of each charging station in a certain area.

[0055] The steps S1 / S2 / S3 / S4 are executed in sequence;

[0056] Optimize the charging infrastructure configuration and establish a transportation-allocation network toll allocation model. In addition to determining the locations of charging facilities, this model can also provide the supply quantity of chargers, providing comprehensive services in terms of location and quantity for electric vehicle charging. This model includes two sub-models, a traffic model and a distribution network model.

[0057] 1) Traffic model:

[0058]

[0059]

[0060] Maximize the coverage of the charging needs of all groups within the planning scope. According to the budget and the maximum number of facilities that can be opened with the available space, ensure that the maximum distance between two consecutive charging stations does not exceed the user's driving mileage L. On the other hand, ensure that the interval between stations does not exceed L. At least one station is required near the origin of L / 2 to meet the route demand, and the capacity of the charging station should be less than or equal to the user's expected charging demand. Setting more budgets and more public spaces will increase the maximum number of chargers installed in a charging station. The minimum number is set according to the minimum charging demand per group per hour and the available capacity of each charging station. This model fully considers the impact of traffic on charging demand management, sets the size and number of charging stations according to traffic mileage, improves the utilization rate of charging stations and users, and innovates the allocation method of charging infrastructure.

[0061] 2) Distribution network model

[0062] In the charging plan, in addition to providing facilities for users, it is also necessary to manage the steady-state and continuous current of the power grid of each power station.

[0063]

[0064] Where: represents the actual power injected at node h in the power grid, and α t represents the hourly electricity price at time t, represents the node power consumption, which is obtained from the basic load and equivalent power load during the charging period. The power load of each charging station is obtained from the number of charging vehicles, SOC, charging rate, battery capacity, and travel distance within each time period. and are the reactive power and injected power load of each node respectively. Equations (12) and (13) control the node voltage in the power grid according to the distribution line data.

[0065] Combine the traffic model with the distribution network model to establish a transportation-allocation network charging allocation model, and determine the optimal locations of charging facilities in the transportation system and the power system.

[0066] This method attempts to minimize the power cost on the power grid under the constraints of power flow distribution. Its purpose is to balance the power flow of the distribution network and reduce the power flow fluctuations during peak hours. Equation (4) indicates that the node power consumption is obtained from the basic load and equivalent power load during the charging period, and the power load of each charging station is obtained from the number of charging vehicles, SOC, charging rate, battery capacity, and travel distance within each time period. Equations (4)-(5) control the node voltage in the power grid according to the distribution line data.

[0067] Combine the traffic model (1)-(3) with the distribution model (4)-(6) to establish a multi-objective model, and determine the optimal locations of charging facilities in the transportation system and the power system. This method effectively allocates the available capacity and prevents over-consumption, especially during peak hours, considering the volatility of demand.

[0068] The growth in the demand for electric vehicles has also stimulated the growth of the market share of electric vehicles among the disabled and the elderly (UDS). The growing demand from UDS requires more accessible charging infrastructure. In addition to incentive programs and charging system layouts, the percentage of UDS users may affect the performance of the system. The more UDS users there are in the system, the more likely it is to change the size of the station and the waiting time of the users. For this reason, charging allocation and incentive programs are applied to networks with different UDS shares in the system. Based on the results, the size of the charging stations will remain unchanged under different UDS percentages in the system. This is because UDS is evenly distributed in the system. However, according to the demographic data of traffic analysis areas, UDS may be more concentrated in specific areas, which increases the demand for FLEX-ACP at some nearby charging stations. In addition to station design, the waiting time in the system may also be affected by the number of UDS. Since the charging system remains unchanged, the waiting time of UDS will increase.

[0069] Therefore, this method proposes Flexible Accessible Charging Points (FLEX-ACP) to improve the accessibility of charging piles for UDS users in charging stations. FLEX-ACP charging piles can switch between "only accessible" and "open to the public" according to the demand status. When it is detected that a UDS user's vehicle is queuing or about to arrive and there is an idle FLEX-ACP charging pile, the system determines that the current demand status is suitable to set the FLEX-ACP charging pile to the "only accessible" mode to give priority to serving UDS users; conversely, if the demand from UDS users is low, while the utilization rate of conventional charging piles is high or there are many queuing vehicles, the system may decide to switch some or all of the FLEX-ACP charging piles to the "open to the public" mode to improve the overall charging efficiency and resource utilization rate to ensure a more flexible and coordinated charging platform for users.

[0070] The transportation-allocation network toll allocation model simulates the behavior of users under incentive programs.

[0071] To encourage users to change their charging behavior, a comprehensive survey was conducted to understand their charging preferences. Based on the survey results, an incentive program was developed to provide rewards to users who choose to charge during congested times or locations. By reducing the utilization rate during peak hours and high-demand locations, a more stable charging rate can be achieved, reducing the charging time of users and thus improving the overall performance of the charging infrastructure.

[0072] This incentive mechanism aims to predict the response of drivers to changes in charging time and location. The value of charging incentives affects users' charging behavior by increasing the level of user utility. The goal of this demand management invention is to reduce the charging time of electric vehicle owners by evenly distributing demand across the network.

[0073]

[0074] By using the level to represent the utility of accepting the behavior of changing the charging fee, that is, the user accepts the incentive for the alternative a ∈ {1, 2, …, A}. On the other hand, ψ0 shows the utility level of the habitual charging behavior. The alternatives include different time periods and potential charging stations, and the user can choose other charging stations other than the current charging station. is the additional utility change generated by switching to the optional a. This value can be positive or negative, depending on the impact of the new decision on the user's charging utility. For example, if by choosing a, the user will switch to a farther station, then the value will be positive. However, if the waiting time at other stations is shorter or the charging rate per hour is lower, choosing these options will increase the charging level of the service, and the corresponding g value is negative.

[0075] In the case of changing the charging behavior, the set of alternatives of each charging station operator and the options they choose highly depend on the utility level of the agent and the service level of the alternatives. When users try to maximize their utility, they will choose the options that meet their needs and offer higher returns among the options. In addition, the agent may not want to change the charging time due to working hours. Therefore, it is expected that users will choose the one with the highest utility value among the choices that are as good as the initial utility value. Although agents tend to maximize their utility, decision-makers have limited resources and are willing to use them to invest in demand management programs to achieve their system goals. Therefore, incentives are provided to a point where it meets the agent's initial utility and they tend to change their charging behavior. The initial benefit level is a benchmark indicating that users will change their behavior in exchange for rewards. This method optimizes the utilization rate of each charging station, reduces the time of charging station congestion during peak periods, fully considers the needs of charging users, and innovates the charging user incentive method.

[0076] Adopt a two-layer optimization method to allocate users to personalized charging options.

[0077] For the optimal station allocation design process, the user will select the charging destination based on the access to charging stations along the route, ignoring the waiting time and availability. First, estimate the reservation payment for the user's current charging choice according to the payment equation. Subsequently, based on the user's route and the charging stations within the route, locate the set of charging stations recommended for the user. According to the flexibility of the user's schedule and the incentives required to change the charging time selection, find the time periods when the user may switch. According to each user's option set and setting the discount range from 0 to 100%, estimate the rewards that each station will offer during each time period. At this stage, use the genetic algorithm to determine the optimal incentive discount. The discount level for each station during each time period is regarded as a variable of the optimization problem. The chromosome will change in each iteration to observe the user behavior based on the capacity and a set of preferred choices. The survival of each candidate is based on minimizing the utilization rate of the system. It should be noted that the comprehensive incentive used to change the user behavior is within the specified budget. Using the optimal incentive set for each station, reallocate the users to find the optimal demand distribution in the next stage. Optimize the network charging demand during the peak hours (7 am to 7 pm), and set the charging cycle every 2 hours. The goal is to reduce the waiting time by allocating the demand among multiple charging stations. In addition, considering the cost-optimized behavior of the users, determine the optimal discount level for each station. This ensures both the efficient management of the charging demand and the satisfaction and convenience of the users.

[0078] The described bilevel optimization problem includes an upper-level problem and a lower-level problem. The upper-level problem attempts to improve the utilization rate of the stations by selecting the best incentive scheme. In addition, at the lower level of the optimization problem, the user attempts to minimize their charging cost by selecting an alternative that maximizes their utility.

[0079] a. Upper-level problem:

[0080]

[0081] Among them, let be the service rate of charger k at time t, represent the number of charger k in charging station h, and use Equation (8) to calculate the number of charging times that the station can serve at t. In the upper-level problem, minimize the maximum demand allocation rate for all stations and the planning period.

[0082] At the highest demand rate among the upper-level objectives, it is necessary to control the stability of the charging server. Reducing the utilization rate will increase the likelihood of available idle chargers when users arrive. The utilization rate of the charging station can be estimated by the number of chargers and the charging rate within the charging station. It should be noted that the arrival rate is initially related to the level of incentives provided. The incentives are reflected in the discount percentage for each station. Therefore, users will receive rewards in the form of a discount on the charging cost. Since the discount that each gas station can offer is limited, the value of the reward cannot exceed the maximum amount.

[0083] This method changes the arrival rate according to the service level of each charging station, thereby reducing the waiting time at the charging station. The charging waiting time is related to the arrival rate and the service level of the waiting. Two new factors for reducing the waiting time are considered.

[0084] The transfer of users and the reduction of the consumption rate will affect the average waiting time for users to access available charging piles. The arrival rate is directly related to the function of the waiting time. However, since this method is dealing with a multi-server non-preemptive priority queuing facility, the arrival rate may have different effects on the waiting time due to the non-linearity in the function. The average waiting time for each station is calculated using the multi-server facility rules. The time-based waiting time for each station demonstrates the impact of the incentive scheme on demand management and the service quality of the charging station. The reduction of the average waiting time indicates that the reduction of the utilization rate is effective and improves the accessibility for users. In addition, the incentive program has successfully reduced the waiting time from more than 20 minutes to less than 6 minutes. In the electric vehicle charging network, minimizing the waiting time is crucial because the charging time itself is time-consuming. Therefore, having an accessible charging infrastructure will make it easier for electric vehicle owners to charge and influence more users to use this mode of transportation.

[0085] b. Lower-level problem

[0086] The lower-level problem is developed to balance user choices based on the decisions of users. Since the charging choices of each user will affect the waiting time, charging rate, and incentives provided by the system, the state of the electric vehicle charging demand management system will be affected. Therefore, the charging choices of other upcoming users will also be affected. On the other hand, users try to minimize the charging cost by finding the most economical charging option. Thus, another optimization problem is formed, which is reflected in the lower-level problem and aims to study the interaction of each user's choice on the electric vehicle charging demand management system and the decisions of other users.

[0087]

[0088] In the lower-level problem, the charging cost in equation (9) represents the charging cost of a user at time t, at charging station h, and with discount x. Each user's chosen alternative must exceed their initial charging cost, and it must be one of the feasible options for each user. Each user's possible alternatives depend on their travel route and flexibility. This method takes into account each user's behavior and the influence among users, reducing the time for users to reach available charging stations and improving the service efficiency of charging stations.

[0089] By studying and evaluating in more depth the users who changed their initial charging choices, 85% of the users changed their charging time, 86% of the users changed their charging location, and 3% of the users maintained their original charging choices. Therefore, it can be seen that a large proportion of users are willing to change their charging behavior according to the incentive scheme, which can also be seen in the survey results. Due to limited user schedules, changing the charging time may not be as smooth as changing the location. By comparing the charging locations with the highest conversion rates, it can be observed that those stations either have high demand or low incentive rates. Therefore, users decide to switch their charging location to other stations with lower demand and correspondingly shorter waiting times. In addition, since other stations are available within their route, this provides an incentive for charging their vehicles, reducing their costs and attracting users to make this mode change.

[0090] The bilevel optimization problem is transformed into a single-level optimization problem by relaxing the objective function of the lower-level problem, and the bilevel optimization problem is transformed into a non-convex MINLP optimization problem.

[0091] First, variable selection is carried out. Only the incentive level x is considered as a variable and is operated as a chromosome in the genetic algorithm. Then, the fitness function is set with the goal of minimizing the maximum demand allocation rate of the system and iterated. Based on the principle of natural selection, the genetic algorithm approximates the optimal solution by continuously evolving the solutions in the population. In each iteration, by changing the chromosome (the value of the incentive level x), the behavioral changes of users based on the charging station capacity and their own preferences are observed. According to the users' choices, the utilization rate of the system is calculated, and the fitness of each candidate solution (chromosome) is evaluated. Solutions with higher fitness are more likely to be retained in the next generation. After multiple iterations, a set of optimal discounts (the values of the incentive level x) that can encourage users to change their charging habits are found.

[0092]

[0093] Equation (10) reflects that this process is an iterative process. Using the objective function as the survival method, through the calculations in this stage, the optimal discount levels that can encourage users to change their charging habits at each station can be obtained;

[0094] In the second stage, the optimal incentive discount obtained in the first stage is used to reallocate users to each charging station to find the optimal demand distribution. During this process, factors such as users' travel routes, charging demands, time flexibility, as well as the capacity and incentives of charging stations are considered. The charging choices of users are adjusted through an optimization algorithm.

[0095] Finally, the branch-and-price-and-cut method is adopted to solve the MINLP problem.

[0096] Solution process of solving the MINLP problem: After determining the user allocation, the problem is transformed into a mixed-integer nonlinear programming (MINLP) problem for solution. The branch-and-price-and-cut algorithm is used to handle the non-convexity and integer variable constraints of the problem during this process. The branching operation divides the solution space of the problem into multiple branches, searches each subspace, and gradually approaches the optimal solution; the price-cutting operation tightens the relaxed solution of the problem and reduces the search space by identifying and utilizing the special structure of the problem, such as valid inequality constraints (cutting planes); the pricing operation involves calculating the lower bound of the optimal solution for each subproblem during the branching process and selecting the most promising branch for further exploration based on the lower bound. By continuously repeating these operations, the optimal solution that satisfies the constraint conditions is finally found, that is, the optimal incentive level x for each charging station at each time period, the charging choices y of users, and the minimum-maximum demand allocation rate z of the system are determined.

[0097] The electric vehicle market has been growing steadily, and adjustments to the transportation network and infrastructure are needed to accommodate the increasing demand and facilitate travel. Although expanding the charging network is a viable option, it also introduces higher power consumption during peak hours, posing challenges to the sustainability of the power grid and power infrastructure. Therefore, formulating a demand management plan is crucial for improving the efficiency and service performance of the charging network.

[0098] Example 1: In this study, the present invention designs an economic optimization method for charging facilities based on electric vehicle charging demand management. It includes the following steps: (1) Considering the overlap of the transportation and power networks in Salt Lake County, determine the charging infrastructure and the capacity of charging stations; (2) Promote social fairness by preferentially providing accessible charging points for users with special needs; (3) Model the payment function based on the personal charging choice utility survey to predict their charging behavior; (4) Establish a two-layer optimization model to optimize the layout of the charging station network and determine the incentives required for each charging station.

[0099] Furthermore, in Step 1, based on the described model, the optimal charging station locations in Salt Lake County, Utah, were found. Assume that the driving range of the electric vehicle is 150 miles and the battery capacity is 30 kWh. Similarly, assume the charging rate is 5 kWh. Assume the market share of EVs is 2%, and 30% of the users belong to the UDS group. Select the electric vehicle market share, assuming the part charged at public charging stations. In addition, the driving range of the sample vehicle is 150 miles and the battery capacity is 30 kWh, which is similar for users. The distribution network used in this method is the IEEE-34 node test feeder. According to the grid and charging demand, the average active load of the feeder is 152 kW, and the active injection load is 55 kW. The reactive loads are 70 kvar and 247 kvar. With the control of the network power load, the minimum node voltage is 0.97 p.u. Before the incentive program, users would select charging stations from locations that intersect their travel routes. However, the discounts offered by the charging stations may affect the users' initial choices.

[0100] Furthermore, in Step 2, the goal of studying the FLEX-ACP charger was to increase the accessibility of UDS. Since UDS has the priority charging right when using FLEX-ACP, it is expected that UDS can reduce the waiting time for charging. For this purpose, the waiting time of UDS in the model proposed in this study was compared with that without FLEX chargers at the stations. The results show that UDS experiences waiting time at all stations. The average waiting time saved by UDS compared with the charging system without FLEX chargers was described. Due to the higher charging demand, the percentage of time saved is the highest at night and at noon. Based on the change in waiting time, it can be concluded that the introduction of FLEX chargers has successfully improved the charging availability of UDS.

[0101] In further Step 3, based on the described model, relevant data in Salt Lake County, Utah, were found for the study. It involves information such as the assumed driving range, battery capacity, charging rate, and market share of electric vehicles, and basic conditions such as considering that some users belong to the UDS group and the number of stations is limited to 10.

[0102] The payment function was constructed based on the results of 112 questionnaires conducted in Salt Lake City. The survey included users' demographic information (such as age, education level, gender, etc.) and daily travel habits (such as travel origin and destination, travel purpose, preferred charging time, etc.). The participants were informed to only consider fast chargers. By providing two stated preferences (SP), the incentives required for each user to change their habitual behavior were found. The alternative options included changing the charging location (station) and time, where the waiting time and the travel time to the station, as non-monetary incentives, would affect the users' decisions. The data collected were resampled according to the US census data of Utah to avoid bias.

[0103] Construct a payment function using the results of logistic regression. Since the decision is a binary variable and the dataset is limited, a quasi-binomial fitting is used for the charging behavior model. The function form is ψ = f(χ, Z, P), where ψ is the payment value, f(.) is the logistic regression model, χ represents the sociodemographic data, Z is the alternative-specific variable including travel time, charging time, and location, and P is the incentive reward for the alternative choice. This function calculates the utility of each charging option for each user, and the utility values for charging time and location are different.

[0104] Calculate the utility of each charging option through ψ = f(χ, Z, P), and then use the logarithmic function to determine the probability of the charging option. The logarithmic function is in the form of

[0105] The option with a higher probability (higher utility function value) is more likely to be selected by the user, thus determining the change in the user's charging behavior. For example, if the user's utility function indicates that a certain alternative charging option is more attractive, then the user has a higher probability of changing the charging location or time. In this way, the payment function modeled based on the survey results of the utility of individual charging options can predict the user's charging behavior.

[0106] In step four, based on information such as the given electric vehicle parameters (such as driving range, battery capacity, etc.), market share, and user demand, determine the optimal charging station location and capacity. By comparing indicators such as station utilization rate, user waiting time, and charging cost under different incentive schemes, the effectiveness of the incentive schemes is verified. Determine the incentives required for each charging station according to the model calculation results. For example, different stations may set different discount incentives, point rewards, etc. according to factors such as their location and demand situation to guide users to choose appropriate stations for charging, thereby optimizing the charging station network layout.

[0107] Relax the objective function of the lower-level problem and transform the bilevel optimization problem into a single-level non-convex mixed-integer nonlinear programming (MINLP) problem for easy solution.

[0108] Use a heuristic algorithm and a branch-and-cut algorithm to solve this problem. The bilevel optimization model is established as follows:

[0109] Upper-level optimization objective: Reduce the station utilization rate by selecting the best incentive scheme. The upper-level model mainly focuses on the overall system objective, that is, how to reasonably distribute the charging demand among stations through incentive measures to avoid overcrowding at some stations and improve the operation efficiency of the entire charging station network.

[0110] Lower - level optimization objective: The lower - level model balances user choices by maximizing its own utility through the charging options selected by users, while considering the users' charging costs and the system's waiting time. That is, from the perspective of individual users, based on the users' utility functions, users will choose charging stations according to their own interests, and the model needs to consider how to achieve system - level optimization while meeting users' needs. In the case study, the charging demand management plan is applied to the Salt Lake County network. Based on information such as given electric vehicle parameters (such as driving range, battery capacity, etc.), market share, and user demand, the optimal charging station locations and capacities are determined.

[0111] Finally, it should be noted that: The above - mentioned embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: They can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; And these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A charging facility economic optimization method based on electric vehicle charging demand management, characterized by: The following steps are involved: Based on the number of charging facilities and charging vehicles at charging stations in a certain area, a transportation-distribution network charging distribution model is established to determine the optimal location of charging facilities in the transportation system and power system, as well as to estimate the capacity and service rate of charging stations; Design incentives to change charging behavior based on the transportation-distribution network charging allocation model to reduce utilization during peak hours and high-demand locations; Construct a two-level optimization problem to distribute user demand among multiple charging stations to reduce waiting time; The double-level optimization problem is transformed into a single-level optimization problem. By relaxing the objective function of the lower-level problem, the double-level optimization problem is transformed into a non-convex MINLP optimization problem. The non-convex MINLP optimization problem is solved to maximize the utilization of the charging demand of each charging station in a certain area.

2. The method for economic optimization of charging facilities based on electric vehicle charging demand management according to claim 1 is characterized in that: The transport-distribution network charging distribution model includes a traffic model and a distribution network model; The expression of the traffic model is as follows: in: Indicates the maximum distance between two consecutive stations, N f represents a set of nodes in the grid that can cover path f, and is sorted by distance. Since electric vehicles have a limited range, the dispersion of charging locations also needs to be considered. f is a binary variable, which takes the value 1 if the path is f. (1) ensures that the maximum distance between two consecutive stations does not exceed the driving range L, (2) ensures that the interval between stations does not exceed L, and (3) requires that at least one station is needed near the L / 2 origin to meet the route requirements; The expression of the distribution network model is as follows: in: represents the actual power injected into the grid node h, α t represents the hourly electricity price at time t, Represents the node power consumption, which is obtained by the basic load and equivalent power load during the charging period. The power load of each charging station is obtained by the number of charging vehicles, SOC, charging rate, battery capacity and travel distance in each time period. and are the reactive power and injected power load of each node respectively. Equation (6) controls the grid node voltage according to the distribution line data. The grid meets the demand by providing consistent power, thereby adjusting the performance of the charging station and realizing steady-state and continuous current management of the grid of each power station.

3. The method for economic optimization of charging facilities based on electric vehicle charging demand management according to claim 1 is characterized in that: The incentive method for changing charging behavior predicts the driver's response to changes in charging time and location. The value of the charging discount affects the user's charging behavior by increasing the user's utility level. The formula expression of the utility level is as follows: The utility of accepting a change in charging behavior is expressed by using the level of usage, That is, the user accepts the incentive for the alternative solution, ψ0 represents the benefit level of habitual charging behavior, It is the additional utility change caused by switching to other optional charging stations. Under this incentive method, the benefits of changing charging behavior are greater than the benefits under habitual behavior.

4. The method for economic optimization of charging facilities based on electric vehicle charging demand management according to claim 1 is characterized in that: The two-level optimization problem includes an upper-level problem and a lower-level problem. The upper-level problem is to improve the utilization rate of the station by selecting the best incentive scheme. The expression of the upper-level problem is as follows: in, is the service rate of charger k at time t, represents the number of chargers k in charging station h, and equation (8) is used to calculate the number of charging times that the station can serve in t; The underlying problem is the interaction of each user's choice on the electric vehicle charging demand management system and other user decisions; Among them, the user's charging cost based on SOC and charging level is C i ,y i,h,t is a binary variable, indicating the choice of charging station h and charging time t by charging user i, x h represents the discount level provided by the station, and formula (10) calculates the minimum charging cost for the user.

5. The method for economic optimization of charging facilities based on electric vehicle charging demand management according to claim 1 is characterized in that: The process of converting the double-level optimization problem into a single-level optimization problem, converting the double-level optimization problem into a non-convex MINLP optimization problem by relaxing the objective function of the lower-level problem, solving the non-convex MINLP optimization problem, and realizing the maximum utilization of the charging demand of each charging station in a certain area is as follows: First, in the first stage, the heuristic algorithm is used as the only variable to solve it; the genetic algorithm is used here, and the specific process is as follows: First, variable selection is performed, considering only the incentive level x as a variable, which is operated as a chromosome in the genetic algorithm; Then, the fitness function is set to minimize the maximum demand allocation rate of the system and iterate. The genetic algorithm is based on the principle of natural selection and approaches the optimal solution by continuously evolving the solutions in the population. In each iteration, by changing the value of the chromosome, that is, the incentive level x, the behavior changes of users based on the capacity of the charging station and their own preferences are observed. According to the user's choice, the utilization rate of the system is calculated, and the fitness of each candidate solution (chromosome) is evaluated. Solutions with high fitness are more likely to be retained to the next generation. After multiple iterations, a set of optimal discount incentive level x values ​​that can encourage users to change their charging habits is found: Formula (10) shows that this process is an iterative process, using the objective function as a survival method. Through the calculation of this stage, the optimal discount level that can encourage users to change their charging habits at each station can be obtained; In the second phase, the optimal incentive discount obtained in the first phase is used to reallocate users to various charging stations to find the optimal demand distribution. In this process, the user's travel route, charging demand, time flexibility, and the capacity and incentives of the charging station are considered, and the user's charging choice is adjusted through the optimization algorithm. Finally, a branch-drop-price approach is adopted to solve the MINLP problem.

6. The method for economic optimization of charging facilities based on electric vehicle charging demand management according to claim 1, characterized in that: The process of using the branch-drop-price approach to solve the MINLP problem is as follows: After determining the user allocation, the problem is transformed into a mixed integer nonlinear programming MINLP problem for solution; in this process, the branch-markdown-pricing algorithm is used to deal with the non-convexity and integer variable constraints of the problem; the branching operation divides the solution space of the problem into multiple branches, searches each subspace, and gradually approaches the optimal solution; the markdown operation tightens the slack solution of the problem and reduces the search space by identifying and utilizing the special structure of the problem, such as effective inequality constraints (cutting planes); the pricing operation involves calculating the lower bound of the optimal solution for each subproblem in the branching process, and selecting the most promising branch for further exploration based on the lower bound. By repeating these operations continuously, the optimal solution that meets the constraints is finally found, that is, the optimal incentive level x for each charging station in each time period, the user's charging choice y, and the minimum and maximum demand allocation rate z of the system are determined.

Images