Planning method of electric vehicle charging station based on CRITIC method and non-cooperative game
The CRITIC-NG method optimizes EV charging station planning by integrating stakeholder interests using the CRITIC method and non-cooperative game theory, addressing conflicts and achieving optimal site selection and capacity configuration for EV charging stations.
Patent Information
- Application Number
- CN202111221924.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-10-20
- Publication Date
- 2024-12-06
- Estimated Expiration
- 2041-10-20
AI Technical Summary
Existing methods for planning electric vehicle (EV) charging stations fail to consider conflicts of interest between charging stations, EV users, and power supply companies, leading to suboptimal planning that does not achieve a win-win situation for all stakeholders.
A two-level planning method using the CRITIC (Criteria Importance Through Intercriteria Correlation) method for indicator weight determination and non-cooperative game theory to analyze the interests of multiple stakeholders, optimizing EV charging station site selection and capacity configuration.
The method improves the accuracy and reliability of EV charging station planning by addressing conflicts and ensuring a win-win situation for charging stations, EV users, and power supply companies through optimal site selection and capacity configuration.
Smart Images

Figure QLYQS_1 
Figure QLYQS_2 
Figure QLYQS_3
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of automobile charging stations, and in particular relates to a planning method for an electric vehicle charging station based on a CRITIC method and a non-cooperative game. Background Art
[0002] In recent years, with the increasing environmental pollution and the gradual depletion of oil resources, vigorously developing and popularizing electric vehicles (EV) has become one of the effective countermeasures, and the planning and construction of EV charging stations are the key links to achieve this countermeasure. Reasonable planning of EV charging stations can not only reduce investment and save land, but also ensure the efficiency of the transportation network and reduce the operating burden of the distribution system. At this stage, with the continuous increase in EV penetration, it is urgent to establish a complete EV charging system.
[0003] In order to make the site layout of EV charging stations more reasonable, the article "Analysis and Evaluation of Infrastructure Configuration Schemes for Electric Vehicle Charging and Swapping Service Network Considering the Acceptance Capacity of Distribution Network" published in the 36(6) issue of "Electric Power Automation Equipment" in 2016 considered multiple factors such as road congestion, service radius, and distribution network acceptance capacity, and established a charging station site selection model through the analytic hierarchy process. Nan Qiqi, Mu Yunfei, Dong Xiaohong, etc. proposed four criteria in "Comprehensive Evaluation Index System and Method for Electric Vehicle Fast Charging Network", namely, fast charging network operation level, user experience, traffic network operation impact, and distribution network operation impact. The analytic hierarchy process was combined with the entropy weight method to calculate the weight of each criterion, forming a charging station site selection index system with a target layer, a criterion layer, and an indicator layer. At the same time, there are also reports abroad that the best location of electric vehicle charging stations in Ankara, Turkey was evaluated using data mining methods, and then the optimal charging station site was selected. A literature proposes a location-aware method for selecting electric vehicle charging station sites. Some people also determine the weights of EV charging station site selection indicators based on the entropy weight method or the hierarchical analysis method. However, the entropy weight method does not consider the conflict of charging demand indicators when assigning weights to indicators. Although the hierarchical analysis method can consider the correlation and conflict between indicators, its indicator weighting is too subjective and the weighting results lack stability.
[0004] In order to provide a reasonable EV charging station capacity configuration plan, Li Hongzhong, Qiang Wei, Gao Yunan, etc. proposed a planning method for electric vehicle charging stations based on user characteristics and traffic information in the article "Charging station planning considering user travel characteristics and available margin of distribution network lines" and Ge Shaoyun, Feng Liang, Liu Hong, etc. proposed a planning method for electric vehicle charging stations based on user characteristics and traffic information in the article "Charging station planning considering traffic information and distribution network capacity constraints" from the perspective of convenient charging for EV users. There are also methods that combine the planning of centralized charging stations with the dispatching of distribution networks, and establish a two-level planning model for centralized charging stations that considers the role of peak shaving and valley filling. Or, considering the impact of user factors and distribution network factors on charging station planning at the same time, a multi-objective planning model for electric vehicle charging stations is established.
[0005] However, although the above-mentioned existing technologies take into account the impact of user factors or distribution network factors on charging station planning, they are all considered from the perspective of the charging station, and are not considered from the perspective of the government level and the whole society. Therefore, when faced with the conflict of interests between charging stations, EV users and power supply companies, multiple stakeholders cannot achieve a win-win situation. Summary of the invention
[0006] In view of this, the present invention proposes a planning method for electric vehicle charging stations based on the CRITIC (Criteria Importance Through Intercriteria Correlation, CRITIC) method and the non-cooperative game (Non-cooperative Game, NG). The method is a two-level planning method for electric vehicle charging stations based on an indicator weight determination method based on indicator correlation and a non-cooperative game, which is abbreviated as the CRITIC-NG planning method. Electric vehicle charging stations are planned through the correlation of indicators and game, which can improve the accuracy and reliability of charging station planning.
[0007] To achieve the above objectives, the technical solution of the present invention is as follows.
[0008] A planning method for electric vehicle charging stations based on the CRITIC method and non-cooperative game, the method comprising the following steps:
[0009] Step 1: Calculate the time and space charging demand of EVs in the planning area;
[0010] Step 2: construct the sub-indicators for EV charging station site selection and weight them using the CRITIC method to establish a comprehensive charging station site selection index that can reflect EV charging demand;
[0011] Step 3: Analyze the interests of the three parties involved in the EV charging station and establish a multi-agent game model of the EV charging station through non-cooperative game theory;
[0012] Step 4. Finally, an upper-level site selection model with the goal of maximizing the comprehensive index of EV charging station site selection and a lower-level capacity determination model with the goal of multi-agent game equilibrium are established. The branch and bound method and the iterative search method with embedded auxiliary objective function are used to iteratively solve the upper and lower layers of the model.
[0013] Specifically,
[0014] In step one, EV charging demand analysis is performed; the charging probability of EVs in the planning area at each time of the day is counted to obtain the time-series charging demand; and various factors affecting EV charging in different partitions are analyzed to obtain the EV spatial charging demand.
[0015] Step 2: Establish an EV charging station site selection index system; consider the traffic factors, functional area factors and power supply reliability factors that affect EV charging, establish corresponding site selection sub-indexes, and determine the weight of each indicator through the CRITIC method, and then establish a comprehensive site selection index for EV charging stations.
[0016] Step three, establish a multi-stakeholder NG model; consider the planning of EV charging stations from the perspective of the government and the whole society, analyze the impact of conflicts of interest among charging stations, EV users and power supply companies on charging station planning decisions, and establish a corresponding NG model.
[0017] Step 4: Establish a two-level planning model for EV charging stations; establish an upper-level site selection model with the goal of maximizing the comprehensive site selection index and a lower-level capacity model with the goal of multi-stakeholder game equilibrium. The upper-level model passes the initial site selection plan to the lower-level model, which optimizes the charging pile capacity based on variables such as the charging load that the charging station needs to meet after site selection, and returns the calculation results to the upper-level model. The upper-level model then optimizes the location and number of charging stations again based on the results of the lower-level model, and iterates until the site selection and capacity solutions for the charging station are optimal.
[0018] The upper model is a mixed integer programming problem, and the present invention adopts the branch and bound method to solve it; the lower model is a game model, and the model may have multiple equilibrium solutions, so the iterative search method with embedded auxiliary objective function is used to determine the optimal Nash equilibrium solution.
[0019] The beneficial effects of the present invention are as follows:
[0020] 1) By analyzing the various factors affecting the site selection of EV charging stations, a corresponding charging station site selection sub-item index system is established, and the CRITIC method is further used to construct a comprehensive charging station site selection index, which overcomes the adverse effects of the conflict and comparison intensity differences among multiple EV charging demand indicators on the reasonable site selection of EV charging stations.
[0021] 2) By adding non-cooperative game links among multiple stakeholders, the optimal configuration of EV charging station capacity is achieved while ensuring a win-win situation for charging stations, power supply companies and EV users. BRIEF DESCRIPTION OF THE DRAWINGS
[0022] Figure 1 It is a basic principle diagram of the CRITIC-NG planning method implemented by the present invention.
[0023] Figure 2 This is the land use nature information map realized by the present invention.
[0024] Figure 3 It is a parking probability map of EV in various functional areas implemented by the present invention.
[0025] Figure 4 This is a diagram of the distribution of intersection nodes implemented by the present invention.
[0026] Figure 5 It is a distribution diagram of the power grid and load points realized by the present invention.
[0027] Figure 6 It is an iterative result diagram of the comprehensive site selection index implemented by the present invention.
[0028] Figure 7 It is a comprehensive index map of site selection in the planning area realized by the present invention.
[0029] Figure 8 This is a site location diagram of the EV charging station implemented by the present invention.
[0030] Fig. 9 This is the location map of the EV charging station under the satellite image implemented by the present invention
[0031] Fig.10 This is a diagram of the access location of the EV charging station implemented by the present invention in the power grid.
[0032] Fig.11 It is a schematic diagram of charging load of charging station 1 and charging station 2 implemented by the present invention.
[0033] Fig.12 It is a schematic diagram of the charging load of the charging station 3 and the charging station 4 implemented in the present invention.
[0034] Fig.13 It is a schematic diagram of the charging load of the charging station 5 and the charging station 6 implemented in the present invention.
[0035] Fig.14 It is a schematic diagram of the charging load of the charging station 7 and the charging station 8 implemented in the present invention. DETAILED DESCRIPTION
[0036] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0037] Figure 1 As shown, the basic principle diagram of the CRITIC-NG planning method implemented by the present invention comprises the following steps:
[0038] Step 1: Calculate the time and space charging demand of EVs in the planning area;
[0039] Step 2: construct the sub-indicators for EV charging station site selection and weight them using the CRITIC method to establish a comprehensive charging station site selection index that can reflect EV charging demand;
[0040] Step 3: Analyze the interests of the three parties involved in the EV charging station and establish a multi-agent game model of the EV charging station through non-cooperative game theory;
[0041] Step 4. Finally, an upper-level site selection model with the goal of maximizing the comprehensive index of EV charging station site selection and a lower-level capacity determination model with the goal of multi-agent game equilibrium are established. The branch and bound method and the iterative search method with embedded auxiliary objective function are used to iteratively solve the upper and lower layers of the model.
[0042] For the analysis of EV charging demand, from a spatial perspective, the spatial distribution of EV charging demand is affected by traffic factors, functional area factors, and power supply reliability factors within the planning area. Different levels of roads will result in different EV traffic. The greater the traffic volume, the greater the demand for EV charging. From the perspective of the functional zoning of the city, residential and commercial areas that can attract more people will undoubtedly have greater EV traffic and will cause more EV charging. The reliability of power supply from the power grid will affect the charging satisfaction of EV users, and EV users will also choose different locations for charging based on the differences in regional power supply reliability. It can be seen that the above factors can affect the charging demand of EVs by affecting the flow of people and vehicles in the planning area, and thus directly affect the site layout of EV charging stations.
[0043] Therefore, the probability function f of the spatial distribution of EV charging demand is constructed s , see formula (1).
[0044] f s =g[Z(x,y),I PE , I EN , I CE ] (1)
[0045] Where: Z(x, y) is the block at the xth column and yth row; IPE It is an indicator of the impact of traffic factors on charging demand; I CE It is an indicator of the impact of functional area factors on charging demand; I EN It is an indicator of how power supply reliability factors affect charging demand.
[0046] In terms of time, the user's travel habits determine the probability of EV charging at each time in a day. Assuming that the EV user's EV battery is fully charged before departure on the day, and charges on the same day after completing the day's travel, after normalizing the statistical data of all EV charging times in the planning area, the maximum likelihood estimation method is used to approximate the EV charging probability at each time to a normal distribution, and its probability density function f t See formula (2).
[0047]
[0048] Where: μ and σ are the expected value and standard deviation of the normal distribution respectively.
[0049] Based on the above analysis of EV spatial charging demand, the present invention quantifies the spatial influencing factors of EV charging into site selection sub-indicators that can reflect EV charging demand, and then uses the CRITIC method to weight and sum each indicator to establish a comprehensive EV charging station site selection index, which is step two.
[0050] Specifically, first, the factors affecting EV charging demand are quantified:
[0051] 1) Traffic factors
[0052] Traffic factors mainly consider the size of traffic flow at different intersections and are used to reflect the actual road capacity. The traffic EV capacity index I PE Assume that the number of road sections connected to the intersection node R is n L , then the traffic flow of intersection node R is shown in formula (3).
[0053]
[0054] Where: R r is the rth intersection node connected to the intersection node R, r = 1, 2, ..., n L ; is the traffic flow of the rth road section connected to the intersection node R in the tth period, vehicles / h.
[0055] 2) Functional area factors
[0056] The functional area factor mainly considers the probability of EV parking in each functional area at different time periods, which is used to reflect the ability of different functional areas to capture EVs. The EV capture ability index I CE The residential area in the planning area is denoted by KH , Industrial area is denoted as K I , commercial area is denoted as K C And other areas are denoted as K O The ability to capture EVs is expressed by the probability of EVs docking in different functional areas. CE,K , see formula (4).
[0057]
[0058] Where: K is each functional area, K = {K H ,K I ,K C ,K O}; is the number of EVs parked in zone K during period t; N EV It refers to the number of EVs in the planning area.
[0059] 3) Power supply reliability factors
[0060] The reliability of power supply of the power grid will affect the charging satisfaction of EV users. EV users will also choose different locations for charging according to the differences in regional power supply reliability. When EV users use the charging pile of a charging station to charge, the higher the frequency of EV charging interruptions caused by power outages and the longer the interruption time, the lower the user's charging satisfaction. When the power supply reliability rate of a load node is greater than or equal to 99.90%, this load node is considered to meet the power supply reliability requirements. Power supply reliability index I EN The specific expression is shown in formula (5).
[0061]
[0062] Where: N EN The number of load nodes that meet the power supply reliability requirements; N N is the total number of load nodes.
[0063] Secondly, for the above indicators, indicator weighting is carried out based on the CRITIC method.
[0064] In order to establish a reasonable EV charging station site selection index system, the present invention needs to weight and sum the EV capacity index, EV capture capacity index and power supply reliability index. However, there are different degrees of correlation and conflict between the site selection sub-item indicators. The degree of correlation between the indicators determines the accuracy and objectivity of the weighting, which cannot be ignored. Therefore, the present invention adopts the CRITIC method to determine the weight of the indicator by judging the amount of information contained in the site selection sub-item indicator. The amount of information contained is measured in two aspects: one is the comparison intensity between the site selection sub-item indicators, which is expressed by the standard deviation. The larger the standard deviation value, the greater the amount of information contained; the second is the conflict between the site selection sub-item indicators, which is expressed by the correlation coefficient between the indicators. The greater the correlation between the two site selection sub-item indicators, the lower the conflict between the two indicators. Then the comparison intensity of the site selection sub-item indicators is multiplied by the conflict quantification value to express the amount of information contained in the indicator, which takes into account both the comparison intensity of the data and the correlation between the site selection sub-item indicators.
[0065] The present invention divides the size of the planned area into n blocks, each block has m site selection sub-item indicators, establishes an n×m indicator matrix A for the entire area, and then uses the CRITIC method to assign weights to each indicator. The CRITIC method weighting steps are as follows:
[0066] Step 1: Standardization of site selection sub-item indicators. The indicator matrix A is shown in formula (6).
[0067]
[0068] Where: a ij is the size of the jth site selection sub-item indicator of the ith block; i = 1, 2, ..., n, n is the number of blocks; j = 1, 2, ..., m, m is the number of indicators.
[0069] The index matrix A is standardized by using the extreme difference method, and the standardized matrix A′ is shown in formula (7).
[0070]
[0071] Step 2: Calculate the correlation coefficient matrix K of the indicators based on the matrix A′. See formula (8).
[0072]
[0073] Where: k hj k is the correlation coefficient between the hth and jth location selection sub-item indexes after standardization; h = 1, 2, ..., m; j = 1, 2, ..., m. hj The calculation formula of is shown in formula (9).
[0074]
[0075] Where: Cov(A h ', A j ') is the hth column vector A in the matrix A' h ' and the jth column vector A j 'The covariance between σ(A h ') is the column vector A h 'Standard deviation; σ(A j ') is the column vector A j ' standard deviation.
[0076] Step 3: Quantify the conflict between the site selection sub-item indicators. The quantitative value b of the conflict between the jth site selection sub-item indicator and other indicators j See formula (10).
[0077]
[0078] Step 4: Calculate the amount of information contained in the indicator. The amount of information is calculated as shown in formula (11).
[0079]
[0080] Where: B j is the amount of information contained in the jth site selection sub-item indicator, B j The larger the value is, the greater the amount of information contained in the j-th site selection sub-indicator, and the greater the relative importance of the indicator.
[0081] Step 5: Calculate the objective weight of the indicator. The objective weight W of the jth site selection sub-item indicator is j See formula (12).
[0082]
[0083] Secondly, the establishment of comprehensive site selection indicators.
[0084] After obtaining the weights of each site selection sub-indicator, the individual numerical values of each indicator are multiplied by the weight of the corresponding indicator and the weighted sum is obtained to obtain the comprehensive site selection index, see formula (13).
[0085]
[0086] Where: is the comprehensive index size of the site selection of the i-th block, i = 1, 2, ..., n; The comprehensive index size of the site selection of the block at the xth column and yth row is used to transform block i and block Z(x,y) according to the sequential counting method so that they are in a corresponding relationship; is the size of the j-th site selection sub-item indicator in the block at the x-th column and y-th row.
[0087] The comprehensive site selection index integrates the traffic factor index, functional area factor index and power supply reliability index that reflect the charging demand, so that it can more comprehensively reflect the charging demand of EVs. When the comprehensive site selection index value of a block is larger, it means that the EV space charging demand of the block is higher, and it will be more reasonable to build an EV charging station here.
[0088] In the operation of EV charging stations, EV charging stations often hope that the capacity of the charging piles configured is reasonable in order to minimize the construction and operation costs and maximize the benefits. EV charging users often hope that the capacity of the charging piles built by the charging station is as large as possible to ensure that their own charging waiting costs are minimized. From the perspective of the power supply company, the more charging piles there are, the more electric vehicles will be connected to the power grid for disorderly charging at the same time, and the greater the probability of charging during peak load periods, the peak-to-valley difference in the power grid load will be further aggravated, and the burden on the power system will be further increased. It can be seen that there is a conflict of interest between EV charging stations, EV users and power supply companies. The three parties are different stakeholders, but they have their own goals and needs. The three parties are closely related and have certain interest connections. The target interests of one party are affected by the strategies of the other two parties. Therefore, the planning decision of EV charging stations is a typical game problem.
[0089] Game theory is a decision-making theory that solves the problem of multiple participants pursuing the maximization of their own interests. The game theory model mainly includes three elements: participants, game strategies, and utility functions. When each game participant has no motivation to change his own strategy or will not change his own strategy alone, the game reaches equilibrium, also known as Nash equilibrium. In the planning decision-making game of EV charging stations, the game strategies of the charging station, EV users and power supply companies are limited. All participants are independent of each other, completely rational, and have mutual understanding of information. The strategy of EV users is manifested in their consideration of charging queue time, and the acceptance degree of EV users to different charging queue times represents multiple strategies of EV users; the strategy of the charging station is the choice of the number and capacity of charging stations by the charging station; the strategy of the power supply company is its acceptance degree of the impact of different numbers and capacities of charging stations connected to the power grid. Therefore, in step three of the present invention, the non-cooperative game theory under a complete information environment is used to establish an EV charging station game theory model, see formula (14).
[0090] M G = {N G ; U A ,U B ,U C ;f A ,f B ,f C} (14)
[0091] Where: N Gis the number of game participants; U is the strategy set; f is the utility function; A, B, and C represent the three parties involved, namely the charging station, EV users, and the power supply company.
[0092] Among them, the utility function of the EV charging station is that the EV charging station is the main body of investment, construction, maintenance and operation of the charging station. In order to maximize its own interests, the EV charging station should ensure the minimum annual cost. Therefore, the utility function of the EV charging station is A The annual payment cost is the minimum. The annual payment cost is composed of annual investment cost, annual operation and maintenance cost and annual profit, see formula (15).
[0093] f A =C qp +C invs -C prs (15)
[0094] Where: C prs is the annual revenue of the charging station, calculated using formula (16); C invs is the annual investment cost of the charging station, calculated using formula (17); C op is the annual operating cost of the charging station, calculated using formula (18).
[0095]
[0096]
[0097]
[0098] Where: N sta is the number of charging stations; is the charging power of EV at the i-th charging station in the t-th period; is the charging electricity price of the EV charging station in the t period; r0 is the discount rate; T is the planning period; is the number of charging piles in the i-th charging station; pc is the rated charging power of the charging pile; C c is the price per unit capacity of the charging pile; is the floor area of the i-th charging station; is the price per unit area of the i-th charging station; is the other investment cost of the i-th charging station; C t b is the electricity price sold by the power grid at time t.
[0099] For the utility function of EV charging users, in order to capture the self-interested charging behavior of EV users in charging stations, the utility function of electric vehicle users is the charging queue cost. EVs arriving at charging stations to receive services are independent of each other and meet the characteristics of stability, no aftereffect, and generality. Therefore, the electric vehicle charging waiting model is reduced to the queuing theory M / M / s model, and the utility function of EV users f B The expression is shown in formula (19).
[0100] f B =365C B T t PN EV (19)
[0101] Where: C B is the waiting cost per unit time; N EV is the number of electric vehicles within the service area of the station; P is the daily charging probability, obtained by formula (2); T t is the average charging queue waiting time, calculated using equations (20) to (23).
[0102]
[0103]
[0104]
[0105]
[0106] Where: t c is the average charging time of a single vehicle; P0 is the probability that all charging piles are idle; ρ is the service intensity of the charging piles; γ is the number of EVs that arrive at the charging station per unit time that satisfies the Poisson flow; ν is the average service rate of the charging piles.
[0107] For the utility function of the power supply company, in addition to the need to build lines and transformers for EV charging stations, the disorderly charging of EVs and charging during peak load periods will increase the burden on the distribution network and cause losses to the power supply company. Therefore, the utility function of the power supply company f C is the line investment, distribution transformer investment and network loss cost, see formula (24).
[0108] f C =C loss +C DT +C line (twenty four)
[0109] Where: C line is the line construction cost connecting the charging station and the substation, calculated using formula (25); C DTThe cost of configuring a dedicated transformer for an EV charging station is calculated using formula (26); C loss is the network loss cost, calculated using formula (27).
[0110]
[0111]
[0112]
[0113] Where: C p is the unit network loss cost; N L The number of feeder lines required to build a charging station; is the network loss of the ith feeder in period t; is the capacity of the transformer of the i-th charging station; C DT is the investment cost of transformer per unit capacity; N B is the number of substations; σ ij is the connection between charging station i and substation j. If a line needs to be built, σ ij =1, otherwise σ ij =0;L ij is the length of the line to be built between substation j and charging station i; C FL is the construction cost per unit length of line.
[0114] Establish a two-level programming model and solve it for EV charging stations.
[0115] First, a two-layer planning model for EV charging stations is established. The planning of EV charging stations includes two aspects: site selection and capacity determination. There are differences between the two, but they are inseparable. Therefore, the present invention models the planning of EV charging stations through a two-layer planning method. The upper layer model is a site selection problem with the goal of maximizing the comprehensive site selection index, and the lower layer is a capacity determination problem with the goal of multi-stakeholder game equilibrium. The upper layer planning passes the site selection plan to the lower layer. The lower layer optimizes the charging pile capacity based on variables such as the service scope of the charging station after site selection and the charging load that needs to be met, and passes the calculation results to the upper layer. Then the upper layer further optimizes the site selection plan based on the power and number of charging piles obtained by the lower layer. Its mathematical model is shown in formula (28).
[0116]
[0117] Where: F1 and F2 are the objective functions of the upper and lower level planning respectively; S sta The capacity of the charging pile required for the charging station is determined by the rated power p of the charging pile. c and the number of charging piles N c Decision; G1 and G2 are the constraints of the upper and lower planning respectively.
[0118] For the upper model:
[0119] 1. Objective Function
[0120] The upper model is the site selection model for EV charging stations, which should comprehensively consider the influence of traffic factors, functional area factors and power grid factors. Therefore, based on the spatial distribution characteristics of EV charging demand, the model selects the location of EV charging stations with the goal of maximizing the comprehensive site selection index. The specific expression is shown in formula (29).
[0121]
[0122] Where: W PE is the weight of the EV capability index; W EN is the weight of the power supply reliability index; W CE To capture the weight of EV capability indicators; They are the size of the passing EV capability index, the size of the power supply reliability index, and the size of the captured EV index of the block in the xth column and the yth row respectively.
[0123] 2. Constraints
[0124] 1) Constraints on the number of charging stations:
[0125]
[0126] Where: The maximum number of charging stations to be built for EVs.
[0127] 2) Constraints on the service coverage of charging station clusters:
[0128] η c ≥1 (31)
[0129] Where: η c It is the service coverage rate of the charging station group.
[0130] Constraints on the construction area of charging stations
[0131]
[0132] Where: s sta [Z(x, y)] is the area of charging stations built in block Z(x, y); The maximum available area for building a charging station in block Z(x,y).
[0133] For the lower model:
[0134] 1. Objective Function
[0135] The lower model is the fixed capacity model of EV charging stations, which is based on game theory. In the capacity planning of EV charging stations, the three participants play games to maximize their own interests, but the optimal solution of the model is determined by the Nash equilibrium of the game. Under this equilibrium strategy, no game participant can obtain greater benefits by changing their own strategy alone. Therefore, it is reasonable to determine the optimal charging pile capacity by the Nash equilibrium of the game. The objective function of the lower model is shown in formula (33).
[0136] minF2={minf A , minf B , minf C} (33)
[0137] 2. Constraints
[0138] 1) Queue waiting time constraints:
[0139]
[0140] Where: T t max The maximum waiting time in the queue
[0141] 2) Single-pile charging power range constraints of charging piles:
[0142]
[0143] Where: They are the maximum and minimum values of the single pile charging power respectively.
[0144] 3) Quantity constraints of charging piles at EV charging stations:
[0145]
[0146] Where: The maximum number of charging posts allowed for a charging station.
[0147] 4) Upper and lower limits of voltage amplitude at distribution network nodes:
[0148]
[0149] Where: V i′ is the voltage amplitude of the distribution network node i′; and are the upper and lower limits of the voltage amplitude of the distribution network nodes respectively; N′ is the load node set of the distribution network in the planning area.
[0150] 5) Distribution line capacity constraints:
[0151] P EVC S +∑Pl、Load ,≤P max (38)
[0152] Where: P WVCS is the total active power consumed by the EV charging station in a day, ∑R l、Load is the active power consumed by other loads on the line during the day, It is the maximum power that the feeder can receive, which is determined by the load on the feeder and the feeder transmission capacity.
[0153] Solution of the bi-level programming model for EV charging stations.
[0154] The upper model is a mixed integer programming problem. The model is relatively simple and can be solved using the branch and bound method.
[0155] The lower-level model is an EV charging station operation game model in which the participants are completely rational and the game strategies are limited. From the existence theorem of Nash equilibrium, it can be seen that the model has a Nash equilibrium solution, but the model is not a unified optimization problem for the optimal solution of the entire system, but a problem in which each participant pursues his own optimal solution. Therefore, the solution method for searching for the global optimal solution is not applicable to the solution of the lower-level model of the present invention.
[0156] In addition, the lower-level game model does not constrain the upper and lower limits of the participants' interests. In the process of solving the model, multiple Nash equilibrium solutions are likely to appear, but the current iterative search method for solving the game model cannot select multiple equilibrium solutions. Therefore, it is necessary to add an auxiliary objective function for selection. In the scenario where the game process is not intervened and only the results of the game are simply evaluated and regulated, the auxiliary objective function is constructed with the minimum loss to the whole society to obtain the optimal Nash equilibrium solution, which can achieve the optimization of interest distribution in a true sense. In view of this, the present invention improves the iterative search solution method, embeds the auxiliary objective function of the equilibrium solution selection into the iterative search method, finds the equilibrium solution of the game model by comparing the interests of the participants under different strategy sets, and then evaluates and selects the results of the game through the auxiliary objective function, thereby determining the optimal equilibrium point of the game model.
[0157] The solution is as follows:
[0158] 1) Construct auxiliary objective function
[0159] Considering that the charging stations to be built should be able to achieve the effect of benefiting the people and promoting business, the auxiliary objective function should be to maximize the interests of the whole society, that is, to minimize the loss of the whole society's interests during the charging station game. The auxiliary objective function is shown in formula (39).
[0160] minF3=min(f A +f B +f C ) (39)
[0161] Where: F3 is the auxiliary objective function.
[0162] 2) Steps to solve the game model;
[0163] Step 1: Initialize the game strategies of the charging station, EV users and power supply companies, and randomly set the equilibrium point.
[0164] Step 2: The three game participants optimize their decisions independently in turn. Based on the results of the previous round of optimization, the participants use genetic algorithms to obtain the optimal combination of the next set of strategies.
[0165] Step 3: Share information about the strategies of the game participants and determine whether the optimal combination meets the constraints. If so, proceed to step 4; if not, return to step 1.
[0166] Step 4: Determine whether the game model can find the Nash equilibrium point under this strategy. If the optimal solutions obtained by each game participant in the adjacent rounds are the same, the game is considered to have reached equilibrium. If the Nash equilibrium requirements are not met, return to step 1 and continue searching.
[0167] Step 5: Record the new equilibrium solution of the game model. Compare this equilibrium solution with the previous equilibrium solution according to formula (39), and retain the equilibrium solution that can reduce the loss of the whole society.
[0168] Step 6: Iterate steps 1 to 5 multiple times until the model cannot find a new Nash equilibrium solution, and finally output the final Nash equilibrium solution.
[0169] Example
[0170] The proposed CRITIC-NG planning method was used to plan a 19.48 km2 area in the economic and technological development zone of a city in Northeast China. 2 EV charging stations are planned in the area. Figure 7 The area enclosed by the lines shown in the figure includes 2 commercial areas, 5 residential areas, 23 industrial areas and 6 other areas. The land use information corresponding to these functional areas is shown in Figure 2. The parking probability of EVs in various functional areas is shown in Figure 3. Figure 3 There are 51 intersection nodes in this area. The distribution of intersection nodes is shown in Figure 4 There are 38 load nodes at the same time. The distribution of power grid and load points in the area is shown in Figure 5There are 2,000 electric vehicles in the planned area, including 400 electric taxis and 1,600 electric private cars. The statistical data of these EV charging times are normalized, and the maximum likelihood estimation method is used to approximate the charging probability of EV at each time to a normal distribution. The parameters of the normal distribution are: μ = 17.3; σ = 3.3. The vehicle parameters are shown in Table 1.
[0171]
[0172] The EV parameters in Table 1 use the CRITIC method to assign weights to the site selection sub-indicators, and the weighting results are shown in Table 2.
[0173] Weight 0.3585 0.3429 0.2986
[0174] Table 2 Index weights determined by CRITIC method
[0175] The branch and bound method is used to solve the upper-level planning model. The results after 150 iterations are as follows: Figure 6 The planned area is divided into 80 square blocks of equal size with a side length of 500m (see Figure 7 ), due to the different functional properties, traffic flow and power grid structure of different blocks, the charging demand of the blocks is different. The comprehensive index of site selection for each block is calculated using formula (13), as follows: Figure 7 The specific values are shown in Table 3.
[0176]
[0177]
[0178] Table 3 Values of comprehensive site selection indicators
[0179] Depend on Figure 7 It can be seen that through the results of the comprehensive site selection indicators, the spatial distribution characteristics of EV charging demand in each block are clear. Per unit area, residential and commercial areas can attract more vehicles and people, which results in greater demand for EV charging; national and provincial roads with strong circulation capacity can attract more EVs, on the contrary, narrow rural roads have less EV traffic; at the same time, the power supply reliability of the power grid also affects the charging location of EVs, and plots with high power supply reliability will attract more EV users to charge.
[0180] 6.2.2 EV Charging Station Site Selection Results
[0181] According to the size of the EV site selection comprehensive index, the EV charging station double-layer planning model was solved to determine the sites of 8 EV charging stations in the planning area. The site locations under the street map are as follows: Figure 8 As shown, the location of EV charging stations in the satellite image is shown in Fig. 9.
[0182] 6.2.3 Capacity Configuration Results of EV Charging Stations
[0183] In the planning game of EV charging stations, the three participants bargained for the capacity required for charging stations at each location. The annual comprehensive cost results of the three participants after the game are shown in Table 4.
[0184]
[0185] Table 4 Annual comprehensive costs of the three participants
[0186] From the analysis in Table 4, in the game process, the construction of fast charging piles can make EV charging stations more profitable and reduce the charging waiting time of EV users, but it will increase the burden on the power grid and increase the losses of power supply companies; when EV users use slow charging piles to charge in residential and industrial areas, although the charging time is long, the charging waiting cost per unit time of EV users is low, so the charging queuing cost of EV users will not be too high.
[0187] Considering the charging environment of different functional areas and the parking probability of EVs, that is, the commercial area has limited land use and high land prices, the scale of EV charging stations that can be built should not be too large, and EVs do not park for a long time in the commercial area, so EV charging stations near the commercial area need to build fast charging piles; on the contrary, industrial and residential areas have no land shortages, large-scale charging stations can be built, and EVs park for a long time in industrial and residential areas, so slow charging piles are suitable. The planning results of EV charging station capacity are shown in Table 5.
[0188]
[0189] Table 5 Planning results of EV charging station capacity
[0190] The rated charging power of the fast charging pile of the present invention is 30kW. An electric car needs to be charged for 30 minutes and an electric taxi needs to be charged for 90 minutes to fully replenish the electric energy. The rated charging power of the slow charging pile is 7kW. An electric car needs to be charged for 120 minutes and an electric taxi needs to be charged for 400 minutes to fully replenish the electric energy.
[0191] EV charging stations are connected to the grid at the following locations: Fig.10 The Monte Carlo simulation was used to simulate the charging process of 2,000 EVs in the planning area. The daily charging demand of the eight EV charging stations after the simulation is shown in Figure 2. Figure 11-14 shown.
[0192] Depend on Figure 7 , Figure 11-14It can be seen that the charging load of EV charging stations near commercial areas is mainly concentrated in 12:00-20:00; the charging load of EV charging stations near residential areas is mainly concentrated in the two periods of 0:00-5:00 and 20:00-24:00; the charging load of EV charging stations near industrial areas is mainly concentrated in 10:00-18:00.
[0193] In summary, the two-level planning method for electric vehicle charging stations based on the CRITIC method and non-cooperative game proposed in this invention has the following two advantages:
[0194] 1) By analyzing the various factors affecting the site selection of EV charging stations, a corresponding charging station site selection sub-item index system is established, and the CRITIC method is further used to construct a comprehensive charging station site selection index, which overcomes the adverse effects of the conflict and comparison intensity differences among multiple EV charging demand indicators on the reasonable site selection of EV charging stations.
[0195] 2) By adding non-cooperative game links among multiple stakeholders, the optimal configuration of EV charging station capacity is achieved while ensuring a win-win situation for charging stations, power supply companies and EV users.
[0196] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A planning method for electric vehicle charging stations based on the CRITIC method and non-cooperative game, the method comprising the following steps: Step 1: Calculate the time and space charging demand of EVs in the planning area; EV charging demand analysis; statistically analyze the charging probability of EVs in the planning area at each time of the day to obtain the time-series charging demand; analyze the various factors affecting EV charging in different partitions to obtain the EV spatial charging demand; Step 2: construct the sub-indicators for EV charging station site selection and weight them using the CRITIC method to establish a comprehensive charging station site selection index that can reflect EV charging demand; Establish an EV charging station site selection index system; consider the traffic factors, functional area factors and power supply reliability factors that affect EV charging, establish corresponding site selection sub-item indicators, and determine the weight of each indicator through the CRITIC method, and then establish a comprehensive site selection index for EV charging stations; Step 3: Analyze the interests of the three parties involved in the EV charging station and establish a multi-agent game model of the EV charging station through non-cooperative game theory; Establish a multi-stakeholder NG model; consider EV charging station planning from the perspective of the government and the whole society, analyze the impact of conflicts of interest between charging stations, EV users and power supply companies on charging station planning decisions, and establish a corresponding NG model; Using non-cooperative game theory under complete information environment, the game theory model of EV charging station is established as shown below: M G ={N G ;U A ,U B ,U C ;f A ,f B ,f C } Where: N G is the number of game participants; U is the strategy set; f is the utility function; A, B, and C represent the three parties involved, namely the charging station, the EV user, and the power supply company; in, The utility function of the EV charging station is as follows: f A =C op +C invs -C prs Where: C prs is the annual revenue of the charging station, C invs is the annual investment cost of the charging station, C op is the annual operating cost of the charging station; The utility function of EV charging users is expressed as follows: f B =365C B T t PN EV Where: C B is the waiting cost per unit time; N EV is the number of electric vehicles within the service area of the station; P is the daily charging probability; T t Average waiting time in charging queue; The utility function of the power supply company is as follows: f C =C loss +C DT +C line Where: C line is the cost of building the line connecting the charging station and the substation, C DT Cost of configuring a dedicated transformer for an EV charging station, C loss is the network loss cost; Step 4: Establish an upper-level site selection model with the goal of maximizing the comprehensive index of EV charging station site selection and a lower-level capacity model with the goal of multi-agent game equilibrium, and use the branch and bound method and the iterative search method with embedded auxiliary objective function to iteratively solve the upper and lower layers of the model; Establish a two-layer planning model for EV charging stations; establish an upper-layer site selection model with the goal of maximizing the comprehensive site selection index and a lower-layer capacity determination model with the goal of multi-stakeholder game equilibrium; the upper-layer model passes the initial site selection plan to the lower-layer, and the lower-layer optimizes the charging pile capacity according to the charging load variables that the charging station needs to meet after the site selection, and returns the calculation results to the upper-layer, and then the upper-layer optimizes the location and number of charging stations again according to the results of the lower layer, and iterates until the site selection and capacity determination plan of the charging station are optimal; For EV charging demand analysis, it is necessary to construct the probability function f of the spatial distribution of EV charging demand s , see the following formula: f s =g[Z(x,y),I PE ,I EN ,I CE ] Where: Z(x, y) is the block at the xth column and yth row; I PE It is an indicator of the impact of traffic factors on charging demand; I CE It is an indicator of the impact of functional area factors on charging demand; I EN It is an indicator of the impact of power supply reliability on charging demand; After normalizing the statistical data of all EV charging times in the planning area, the maximum likelihood estimation method is used to approximate the charging probability of EVs at each time to a normal distribution, and its probability density function f t See the following formula: Where: μ and σ are the expected value and standard deviation of the normal distribution respectively; The indicator weighting based on the CRITIC method is to weight and sum the EV capacity index, EV capture capacity index and power supply reliability index. However, there are different degrees of correlation and conflict between the site selection sub-item indicators. The size of the planned area is divided into n blocks, each block has m site selection sub-item indicators, and an n×m indicator matrix A is established for the entire area. Then, the CRITIC method is used to assign weights to each indicator. The CRITIC method weighting steps are as follows: Step 1: Standardize the site selection sub-item indicators. The indicator matrix A is as follows: Where: a ij is the size of the jth site selection sub-item indicator of the i-th block; i = 1, 2, ..., n, n is the number of blocks; j = 1, 2, ..., m, m is the number of indicators; The index matrix A is standardized by the extreme difference method, and the standardized matrix A′ is shown in the following formula: Step 2: Calculate the correlation coefficient matrix K of the indicators based on the matrix A′, as shown in the following formula: Where: k hj is the correlation coefficient between the hth and jth location selection sub-item indexes after standardization; h = 1, 2, ..., m; j = 1, 2, ..., m, k hj The calculation formula is as follows: Where: Cov(A h ', A j ') is the hth column vector A in the matrix A' h ' and the jth column vector A j 'The covariance between σ(A h ') is the column vector A h 'Standard deviation; σ(A j ') is the column vector A j 'Standard deviation; Step 3: Quantify the conflict between the site selection sub-item indicators. The conflict quantification value b between the jth site selection sub-item indicator and other indicators j See the following formula: Step 4: Calculate the amount of information contained in the indicator. The information calculation is as follows: Where: B j is the amount of information contained in the jth site selection sub-item indicator, B j The larger the value is, the greater the amount of information contained in the jth site selection sub-index, and the greater the relative importance of the index; Step 5: Calculate the objective weight of the indicator. The objective weight of the j-th site selection sub-item indicator W j See the following formula: Establishment of comprehensive site selection indicators: After obtaining the weights of each site selection sub-indicator, multiply the individual values of each indicator by the weight of the corresponding indicator for weighted summation to obtain the comprehensive site selection index, as shown in the following formula: Where: is the comprehensive index size of the site selection of the i-th block, i = 1, 2, ..., n; The comprehensive index size of the site selection of the block at the xth column and yth row is used to transform block i and block Z(x,y) according to the sequential counting method so that they are in a corresponding relationship; is the size of the jth site selection sub-item index of the block at the xth column and yth row; Among them, the two-level planning model of EV charging station is established and solved as follows: First, a two-layer planning model for EV charging stations is established. The upper model is a site selection problem with the maximum comprehensive site selection index as the goal, and the lower model is a capacity problem with the goal of multi-stakeholder game equilibrium. The upper planning passes the site selection plan to the lower layer, and the lower layer optimizes the charging pile capacity according to the service scope of the charging station after site selection and the charging load variables that need to be met, and passes the calculation results to the upper layer. Then the upper layer further optimizes the site selection plan according to the power and number of charging piles obtained by the lower layer. The mathematical model is shown in the following formula: Where: F1 and F2 are the objective functions of the upper and lower level planning respectively; S sta The capacity of the charging pile required for the charging station is determined by the rated power p of the charging pile. c and the number of charging piles N c Determination; G1 and G2 are the constraints of the upper and lower planning respectively; The two-level planning model of EV charging station is solved as follows: 1) Construct auxiliary objective function The auxiliary objective function is as follows: minF3=min(f A +f B +f C ) Where: F3 is the auxiliary objective function; 2) Steps to solve the game model; Step 1: Initialize the game strategies of the charging station, EV users and power supply companies, and randomly set the equilibrium point; Step 2: The three game participants optimize their decisions independently in turn. Based on the results of the previous round of optimization, the participants use genetic algorithms to obtain the optimal combination of the next set of strategies. Step 3: Share the strategies of the game participants and determine whether the optimal combination meets the constraints. If so, proceed to step 4. If not, return to step 1. Step 4: Determine whether the game model can find the Nash equilibrium point under this strategy. If the optimal solutions obtained by each game participant in the adjacent rounds are the same, it is considered that the game has reached equilibrium at this time; if it does not meet the Nash equilibrium requirements, return to step 1 and continue searching; Step 5: Record the new equilibrium solution of the game model, compare this equilibrium solution with the previous equilibrium solution according to the auxiliary objective function formula, and retain the equilibrium solution that can reduce the loss of the whole society; Step 6: Iterate steps 1 to 5 multiple times until the model cannot find a new Nash equilibrium solution, and finally output the final Nash equilibrium solution.
Citation Information
Cited By
A low-carbon public charging facility site selection method
CN122966293A
