Charging station planning method and system based on rural typical scene division, computer equipment and readable storage medium

Abstract

The invention discloses a charging station planning method and system based on rural typical scene division, computer equipment and a readable storage medium, and the method comprises the steps: building a double-layer charging station planning model based on an upper objective function and constraint conditions and a lower objective function and constraint conditions; the upper-layer target function aims at minimizing the investment cost of a constructor, and the lower-layer target function aims at minimizing the charging cost of a user; solving the double-layer charging station planning model, predicting the charging load of the charging station through the prediction model according to the target rural basic data in the solving process, and obtaining a charging station configuration point of a total optimal solution of an upper-layer target function and a lower-layer target function; and configuring the charging station in the rural area according to the charging station configuration point of the optimal solution. A rural charging station planning architecture of a double-layer charging station planning model based on a cost optimization target of a constructor and a user charging cost minimization target is combined with corresponding constraint conditions to realize double-layer comprehensive consideration of economy and practicability of rural charging station planning.

Description

Technical Field

[0001] The present invention relates to charging station planning, and in particular to a charging station planning method, system, computer equipment and readable storage medium based on the division of typical rural scenarios. Background Art

[0002] With the booming electric vehicle industry, the number of electric vehicles in rural areas has been increasing year by year. Rural electric vehicles are becoming a new force in rural transportation, with applications spanning daily commuting, agricultural production and transportation, and rural logistics and distribution. They play a vital role in improving rural transportation convenience and energy efficiency. However, the development of supporting rural charging stations has lagged behind, becoming a key factor hindering the further adoption of electric vehicles in rural areas. Consequently, the construction of rural charging stations faces numerous challenges. Firstly, rural power grid infrastructure is relatively weak, and power supply is unstable, making it difficult to meet the power capacity and quality requirements for large-scale charging station construction and operation. Secondly, rural areas are vast, with villages scattered across a wide area, and user charging needs are dispersed, diverse, and uncertain, making the site selection and layout planning of charging stations extremely challenging.

[0003] The current planning process for rural charging stations presents numerous pressing challenges. Traditional planning methods often lack careful consideration of the complex rural geography and regional differences in functionality. They fail to accurately capture the charging behavior and demand characteristics of users in different regions, resulting in irrational charging station layouts and inefficient resource allocation. Summary of the Invention

[0004] Purpose of the invention: In view of the above shortcomings, the present invention provides a charging station planning method, system, computer equipment and readable storage medium based on the division of typical rural scenarios to improve the rationality of the capacity configuration and construction scale determination of charging stations.

[0005] Technical solution: To solve the above problems, the present invention adopts a charging station planning method based on the division of typical rural scenarios, which includes the following steps:

[0006] A two-layer charging station planning model is constructed based on an upper-layer objective function and upper-layer constraints, as well as a lower-layer objective function and lower-layer constraints. The upper-layer objective function aims to minimize the builder's investment cost, and the upper-layer constraints include the power constraint and the area constraint of the charging station. The lower-layer objective function aims to minimize the user's charging cost, and the lower-layer constraints include the remaining power constraint of the electric vehicle and the waiting time constraint in the queue.

[0007] Determine the target rural type, obtain basic data of the target rural area, and solve the two-layer charging station planning model based on the basic data of the target rural area. During the solution process, the charging load of the charging station is predicted through a prediction model based on the basic data of the target rural area to obtain the overall optimal solution of the upper and lower objective functions; the overall optimal solution of the upper and lower objective functions is used to represent the optimal configuration point planning scheme for configuring charging stations in rural areas.

[0008] Furthermore, the upper layer objective function is:

[0009] min f upp =C 1,θ +C 2,θ

[0010] Among them, C 1,θ represents the construction cost; C 2,θ represents operating costs;

[0011] Construction cost C 1,θ and operating costs C 2,θ The calculation formula is:

[0012]

[0013] in, represents the area occupied by charging station i, Represents the charging station capacity of charging station i, i=1,2,3....N CS , N CS is the number of charging stations, a θ represents the unit land price at the construction location of charging station i under the rural type θ, b θ represents the unit cost of the charging pile and related supporting equipment of the construction unit under the rural type, c θ represents the cost of distribution transformers and cables under the rural type; k1 and k2 represent cost conversion coefficients, μ represents the equipment maintenance rate; y represents the operating life of the charging station;

[0014] The upper-level constraints include:

[0015] Power constraints of charging stations:

[0016]

[0017] Among them, P i,j represents the output power of the jth charging device in charging station i, represents the number of charging devices in charging station i; P i,max,θ is the maximum output power of charging station i under θ rural type;

[0018] Charging station floor space constraints:

[0019]

[0020] Among them, S i,j is the area occupied by the jth charging equipment in charging station i; S i,max is the maximum floor space of charging station i.

[0021] Furthermore, the lower layer objective function is:

[0022] minf low =C EV

[0023] Among them, C EV Charging costs for electric vehicle users;

[0024] Charging cost C for electric vehicle users EV The calculation formula is:

[0025]

[0026] Among them, C u,t The cost of electric energy lost when driving an electric vehicle to a charging station; The equivalent time loss cost required for electric vehicle users to travel to charging stations; is the necessary queuing time cost faced by electric vehicle users after driving to the target charging station; M represents the number of all electric vehicle users in the planning area;

[0027] Electricity loss cost C u,t The calculation formula is:

[0028]

[0029] Among them, D uj is the distance between electric vehicle user u and the best charging station; L u is the total cruising range of the electric vehicle u; C u is the battery capacity of the electric vehicle u; η is the vehicle charging and discharging efficiency coefficient; C cos is the unit electric energy loss cost of electric vehicle u;

[0030] Equivalent time loss cost The calculation formula is:

[0031]

[0032] Among them, λ tc is the loss cost per unit time; V u is the speed of the electric vehicle;

[0033] Necessary queuing time cost The calculation formula is:

[0034]

[0035] Among them, T u Waiting time for charging:

[0036] The lower level constraints include:

[0037] Remaining power constraints of electric vehicles:

[0038] S u ×L u ≥D uj

[0039] Among them, S u The remaining power of the electric vehicle;

[0040] Queue waiting time constraints:

[0041]

[0042] Among them, T u Waiting time for electric vehicle users, The maximum waiting time in queue for electric vehicle users.

[0043] Furthermore, the target rural types include tourist villages, sparsely populated villages, and villages with service areas; the judging indicators are: the presence or absence of service areas, the tourism heat index K1, and the residential dispersion index K2;

[0044] When a rural area has no service area and the tourism heat index K1 exceeds a specific threshold, it is judged as a tourist village; when a rural area has no service area and the tourism heat index does not exceed the specific threshold K1, and the resident residential dispersion index K2 exceeds a specific threshold, it is judged as a sparse village; when a rural area has a service area, it is judged as a village with a service area.

[0045] Furthermore, the prediction model adopts an electric vehicle load prediction model based on the GAT-TCN hybrid algorithm. First, the rural coupling network is represented by a graph structure. The spatial features of the graph are learned through the GAT model. The extracted spatial features are input into the TCN model. The TCN model learns the temporal features of the graph features and predicts the charging load of each charging station in different time periods.

[0046] The graph feature is: X=[X EV ,X CS ,X TN ,X PN ]

[0047]

[0048] Among them, X EV represents the electric vehicle node characteristics, X CS represents the charging station node characteristics, X TN Traffic node characteristics, X PN Distribution network node characteristics, t represents time, l u Indicates the position of the electric vehicle, S u Indicates the SOC of electric vehicle battery, C u Indicates the battery capacity of electric vehicles, Indicates the user's desired state of charge. Indicates the state of charge when the user arrives at the charging station. represents the speed of the electric vehicle, represents the planned parking time of electric vehicle users, Indicates charging completion time. Indicates the time when the user arrives at the charging station; s CS Indicates the number of charging piles in the charging station, L CS Indicates the location of the charging station, Indicates charging power, C cos,t Indicates the charging electricity price; l TN represents the length of the road segment, ν TN Indicates the passing speed, g TN represents the traffic volume; P PN Indicates the active power of the distribution network node, Q PN Represents the reactive power of the distribution network node, U PN Indicates the voltage at the distribution network node.

[0049] Furthermore, the specific steps of solving the double-deck charging station planning model include:

[0050] Initialize the parameters of the double-deck charging station planning model;

[0051] Generate an initial particle swarm, where each particle represents the location of a charging station;

[0052] Calculate the value of the upper layer objective function and the value of the lower layer objective function for each particle, and take the weighted sum of the value of the upper layer objective function and the value of the lower layer objective function as the fitness value; find the particle with the best fitness value in the entire particle swarm;

[0053] The particles are iteratively updated, and the particle with the best fitness value is output at the end of the iteration to obtain the optimal solution.

[0054] The present invention also adopts a charging station planning system based on the division of typical rural scenarios, including:

[0055] A model construction module is used to construct a two-layer charging station planning model based on an upper-layer objective function and upper-layer constraints, and a lower-layer objective function and lower-layer constraints; the upper-layer objective function aims to minimize the builder's investment cost, and the upper-layer constraints include the power constraint and the charging station footprint constraint; the lower-layer objective function aims to minimize the user's charging cost, and the lower-layer constraints include the remaining power constraint of the electric vehicle and the queue waiting time constraint;

[0056] The solution module is used to determine the target rural type, obtain the basic data of the target rural area, and solve the two-layer charging station planning model based on the basic data of the target rural area. During the solution process, the charging load of the charging station is predicted through the prediction model based on the basic data of the target rural area to obtain the overall optimal solution of the upper and lower objective functions; the overall optimal solution of the upper and lower objective functions is used to represent the optimal configuration point planning scheme for configuring charging stations in rural areas.

[0057] Furthermore, the upper layer objective function is:

[0058] minf upp =C 1,θ +C 2,θ

[0059] Among them, C 1,θ represents the construction cost; C 2,θ represents operating costs;

[0060] Construction cost C 1,θ and operating costs C 2,θ The calculation formula is:

[0061]

[0062] in, represents the area occupied by charging station i, Represents the charging station capacity of charging station i, i=1,2,3....N CS , N CS is the number of charging stations, a θ represents the unit land price at the construction location of charging station i under the rural type θ, b θ represents the unit cost of the charging pile and related supporting equipment of the construction unit under the rural type, c θ represents the cost of distribution transformers and cables under the rural type; k1 and k2 represent cost conversion coefficients, μ represents the equipment maintenance rate; y represents the operating life of the charging station

[0063] The upper-level constraints include:

[0064] Power constraints of charging stations:

[0065]

[0066] Among them, P i,j represents the output power of the jth charging device in charging station i, represents the number of charging devices in charging station i; P i,max,θ is the maximum output power of charging station i under θ rural type;

[0067] Charging station floor space constraints:

[0068]

[0069] Among them, S i,j is the area occupied by the jth charging equipment in charging station i; S i,max is the maximum floor space of charging station i.

[0070] Furthermore, the lower layer objective function is:

[0071] minf low =C EV

[0072] Among them, C EV Charging costs for electric vehicle users;

[0073] Charging cost C for electric vehicle users EV The calculation formula is:

[0074]

[0075] Among them, C u,t The cost of electric energy lost when driving an electric vehicle to a charging station; The equivalent time loss cost required for electric vehicle users to travel to charging stations; is the necessary queuing time cost faced by electric vehicle users after driving to the target charging station; M represents the number of all electric vehicle users in the planning area;

[0076] Electricity loss cost C u,t The calculation formula is:

[0077]

[0078] Among them, D uj is the distance between electric vehicle user u and the best charging station; L u is the total cruising range of the electric vehicle u; C u is the battery capacity of the electric vehicle u; η is the vehicle charging and discharging efficiency coefficient; C cos is the unit electric energy loss cost of electric vehicle u;

[0079] Equivalent time loss cost The calculation formula is:

[0080]

[0081] Among them, λ tc is the loss cost per unit time; V u is the speed of the electric vehicle;

[0082] Necessary queuing time cost The calculation formula is:

[0083]

[0084] Among them, T u Waiting time for charging:

[0085] The lower level constraints include:

[0086] Remaining power constraints of electric vehicles:

[0087] S u ×L u ≥D uj

[0088] Among them, S u The remaining power of the electric vehicle;

[0089] Queue waiting time constraints:

[0090]

[0091] Among them, T u Waiting time for electric vehicle users, The maximum waiting time in queue for electric vehicle users.

[0092] Furthermore, the target rural types include tourist villages, sparsely populated villages, and villages with service areas; the judging indicators are: the presence or absence of service areas, the tourism heat index K1, and the residential dispersion index K2;

[0093] When a rural area has no service area and the tourism heat index K1 exceeds a specific threshold, it is judged as a tourist village; when a rural area has no service area and the tourism heat index does not exceed the specific threshold K1, and the resident residential dispersion index K2 exceeds a specific threshold, it is judged as a sparse village; when a rural area has a service area, it is judged as a village with a service area.

[0094] Furthermore, the prediction model adopts an electric vehicle load prediction model based on the GAT-TCN hybrid algorithm. First, the rural coupling network is represented by a graph structure. The spatial features of the graph are learned through the GAT model. The extracted spatial features are input into the TCN model. The TCN model learns the temporal features of the graph features and predicts the charging load of each charging station in different time periods.

[0095] The graph feature is: X=[X EV ,X CS ,X TN ,X PN ]

[0096]

[0097] Among them, X EV represents the electric vehicle node characteristics, X CS represents the charging station node characteristics, X TN Traffic node characteristics, X PN Distribution network node characteristics, t represents time, l u Indicates the position of the electric vehicle, S u Indicates the SOC of electric vehicle battery, C u Indicates the battery capacity of electric vehicles, Indicates the user's desired state of charge. Indicates the state of charge when the user arrives at the charging station. represents the speed of the electric vehicle, represents the planned parking time of electric vehicle users, Indicates charging completion time. Indicates the time when the user arrives at the charging station; s CS Indicates the number of charging piles in the charging station, L CS Indicates the location of the charging station, Indicates charging power, C cos,t Indicates the charging electricity price; l TN represents the length of the road segment, ν TN Indicates the travel speed, g TN represents the traffic volume; P PN Indicates the active power of the distribution network node, Q PN Represents the reactive power of the distribution network node, U PN Indicates the voltage at the distribution network node.

[0098] Furthermore, the specific steps of solving the double-deck charging station planning model include:

[0099] Initialize the parameters of the double-deck charging station planning model;

[0100] Generate an initial particle swarm, where each particle represents the location of a charging station;

[0101] Calculate the value of the upper layer objective function and the value of the lower layer objective function for each particle, and take the weighted sum of the value of the upper layer objective function and the value of the lower layer objective function as the fitness value; find the particle with the best fitness value in the entire particle swarm;

[0102] The particles are iteratively updated, and the particle with the best fitness value is output at the end of the iteration to obtain the optimal solution.

[0103] The present invention also adopts a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program.

[0104] The present invention also adopts a computer-readable storage medium having a computer program stored thereon, and the computer program implements the steps of the above method when executed by a processor.

[0105] Beneficial effect: Compared with the existing technology, the significant advantage of the present invention is that the present invention proposes a rural charging station planning framework based on a two-layer charging station planning model. By constructing the cost optimization goal of the upper-level charging station builders and the charging cost minimization goal of the lower-level users, combined with the corresponding constraints, the present invention achieves a comprehensive consideration of the economic and practical aspects of rural charging station planning.

[0106] This paper establishes a charging load forecasting model based on the GAT-TCN hybrid algorithm. It uses GAT and TCN to extract the spatiotemporal characteristics of multi-node coupled networks such as charging stations, transportation, electric vehicles, and distribution networks in complex rural environments. Compared with traditional forecasting models, it greatly improves the forecasting accuracy and can better adapt to the dynamic changes in rural charging demand.

[0107] The present invention constructs an intelligent rural charging station planning system, whose data acquisition and processing module can effectively integrate multi-source heterogeneous data; the charging load prediction module realizes accurate load prediction; the charging station layout optimization module solves the planning model to obtain the optimal layout; the constraint condition management module ensures the stable operation of the power system; and the system visualization module provides intuitive quantitative evaluation basis and interactive demonstration function. BRIEF DESCRIPTION OF THE DRAWINGS

[0108] Figure 1 Schematic diagram of the flow of the charging station planning method in the present invention.

[0109] Figure 2 This is a framework diagram of the charging station planning method in the present invention.

[0110] Figure 3 This is the rural distribution network topology diagram based on IEEE-123 nodes in the present invention.

[0111] Figure 4Schematic diagram of load prediction accuracy of the GAT-TCN hybrid algorithm model in the present invention.

[0112] Figure 5 This is a schematic diagram of the typical daily charging load demand prediction results of the three functional area villages in the present invention.

[0113] Figure 6 This is a schematic diagram of the relationship between the specific costs and the number of charging stations built in the present invention.

[0114] Figure 7 This is a schematic diagram of the location results of the optimal rural charging station planning in the present invention.

[0115] Figure 8 This is a structural block diagram of the charging station planning system in the present invention. DETAILED DESCRIPTION

[0116] Example 1

[0117] like Figure 1 As shown, in this embodiment, a charging station planning method based on the division of typical rural scenarios is proposed to achieve the optimal layout of rural charging station site selection and construction, and a rural charging station planning framework based on a graph spatiotemporal network prediction model and a two-layer charging station planning model is proposed. The following steps are included: based on the upper objective function and upper constraint conditions and the lower objective function and lower constraint conditions, a two-layer charging station planning model is constructed; the upper objective function aims to minimize the builder's investment cost, and the upper constraint conditions include the power constraint of the charging station and the area constraint of the charging station; the lower objective function aims to minimize the user's charging cost, and the lower constraint conditions include the remaining power constraint of the electric vehicle and the queue waiting time constraint;

[0118] Determine the target rural type, obtain basic data of the target rural area, and solve the two-layer charging station planning model based on the basic data of the target rural area. During the solution process, the charging load of the charging station is predicted through a prediction model based on the basic data of the target rural area to obtain the overall optimal solution of the upper and lower objective functions; the overall optimal solution of the upper and lower objective functions is used to represent the optimal configuration point planning scheme for configuring charging stations in rural areas.

[0119] S1: Based on the upper-level objective function and upper-level constraints and the lower-level objective function and lower-level constraints, a two-layer charging station planning model is constructed, specifically including:

[0120] S1.1 Upper-level charging station planning model

[0121] 1) Upper-level objective function

[0122] As the main investor, the constructor of the charging station aims to maximize the profitability of the charging station or minimize its investment cost. The investment cost of the charging station mainly includes two categories: construction cost C1 and operation cost C2. The upper-level planning objective function is:

[0123] min f upp =C 1,θ +C 2,θ (1)

[0124] Construction cost C 1,θ It is mainly composed of land costs and the costs of charging facilities and related supporting equipment. The land cost depends on the area occupied by the charging station, while the cost of charging facilities and equipment is related to the capacity of the charging station. θ is the rural type identifier, θ∈{tourist village, sparse village, service village}. The construction cost is:

[0125]

[0126] Where: and Represent the area occupied by charging station i and the capacity of the charging station, i=1,2,3....N CS , N CS is the number of charging stations; a θ 、b θ 、c θ They represent the unit land price at the construction location of charging station i under the three types of rural areas, the unit cost of the construction unit charging pile and related supporting equipment, and the cost of equipment such as distribution transformers and cables.

[0127] The operating cost is composed of the maintenance cost and service cost of all operating facilities in the charging station. The larger the capacity of charging station i, the more users it serves, and the higher the corresponding service cost. The operating cost is:

[0128]

[0129] Where: k1 and k2 are cost conversion coefficients; μ is the equipment maintenance rate; y is the operating life of the charging station.

[0130] 2) Constraints

[0131] The upper-level planning must meet the power limit, floor space limit, and power quality requirements of the charging station:

[0132]

[0133] Where: P i,j is the output power of the jth charging device in charging station i, is the number of charging devices in charging station i; P i,max,θis the maximum output power of charging station i under the rural type θ.

[0134] Charging station floor space restrictions:

[0135]

[0136] Where: S i,j is the area occupied by the jth charging equipment in charging station i; S i,max is the maximum floor space of charging station i.

[0137] S1.2 Lower-level charging station planning model

[0138] 1) Objective function

[0139] Under the goal of minimizing the investment cost of the upper-level builders, there is an optimal location and capacity for charging station i. Users choose appropriate charging locations to minimize the charging costs of rural residents. Taking into account the travel behavior, charging preferences, and charging costs of electric vehicle users, the objective function of the lower-level planning is:

[0140] minf low =C EV (6)

[0141] Where: C EV Charging costs for electric vehicle users.

[0142] The charging cost for electric vehicle users mainly includes time cost and power loss cost:

[0143]

[0144] Where: C u,t The cost of electric energy lost when driving an electric vehicle to a charging station; The equivalent time loss cost required for electric vehicle users to travel to charging stations; is the necessary queuing time cost faced by electric vehicle users after driving to the target charging station; M represents the number of all electric vehicle users in the planning area.

[0145]

[0146] Where: D uj is the distance between electric vehicle user u and the best charging station; L u is the total cruising range of the electric vehicle u; η is the vehicle charging and discharging efficiency coefficient; C u 、C cos They are the battery capacity of electric vehicle u and the unit energy loss cost.

[0147]

[0148] Where: tc is the loss cost per unit time; T u is the waiting time required for charging; V u is the speed of the electric vehicle.

[0149] 2) Constraints

[0150] The remaining power of an electric vehicle must be greater than the power consumed when driving to the charging station.

[0151] S u ×L u ≥D uj (11)

[0152] Where: S u The remaining power of the electric vehicle.

[0153] The queue waiting time constraint is:

[0154]

[0155] Where: The maximum waiting time in queue for electric vehicle users.

[0156] S2: Determine the rural type and establish a charging load prediction model based on the GAT-TCN hybrid algorithm.

[0157] Rural functional zoning can be categorized into three typical types: tourist villages, sparsely populated villages, and villages with service areas. The unit land price for charging station construction, the unit cost of charging piles and related supporting equipment, and the cost of distribution transformers and cables vary across different village types, impacting the investment costs of builders.

[0158] Tourist Village: The core area consists of tourist attractions and visitor reception areas. During peak season, demand for charging electric tour buses continues even when the scenic area is open, and peak demand for parking lots and other facilities coincides with tourist hours. In the tourist reception area, hotels and guesthouses experience demand for electric vehicles from tourists at night and early morning, and this demand is significantly affected by peak and off-peak seasons. For tourist villages, a tourism heat index, K1, can be defined to measure the intensity of tourism activity. The calculation formula is:

[0159]

[0160] Where: V i represents the number of tourists on day i, n is the number of days in the statistical period, and D is the area of ​​the village. If there is no service area and the tourism heat index K1>K exceeds a certain threshold, the area can be determined as a tourist village.

[0161] Sparse villages: Residential areas are scattered, and residents have little and scattered demand for electric vehicle charging. There is demand at various points after returning home at night. In sparse villages, the residential dispersion index K2 can be used to describe the characteristics of residential areas. The calculation formula is:

[0162]

[0163] Where: d k,k+1 is the distance between the kth household and the k+1th household, and p is the number of households. If there is no service area, the tourism heat index K1≤K, and the residential dispersion index K2>100, the village is judged to be sparsely populated.

[0164] Villages with service areas: If there is a highway service area, it will be directly determined as a service area village. The demand for ordinary highway service areas increases significantly before and after traffic peaks.

[0165] Structural representation of rural coupling network graph:

[0166] The adjacency matrix represents the connections between nodes, and its selection has a great impact on the performance of the model. Let the adjacency matrix be A∈R N×N , N is the rural coupling network G RN The total number of nodes, including charging station nodes, traffic nodes, electric vehicle nodes, and distribution network nodes, where a ij Represents the relationship between nodes i and j, i, j = 1, 2, ..., N. When a ij = 0, it means that there is no spatial relationship between the two nodes. The method of constructing the graph mainly depends on the geographical distance between charging station nodes, traffic nodes, electric vehicle nodes, and distribution network nodes or the correlation between them. RN =(A,X) topological model is as follows:

[0167]

[0168] X∈R N×V is a feature matrix used to store the energy and information interaction characteristics of the environment. V represents the number of features of each node. The feature vectors of each subject are concatenated to obtain X = [X EV ,X CS ,X TN ,X PN ].

[0169]

[0170] Where: t, l u ,S u ,C u , They represent time, electric vehicle location, electric vehicle battery SOC, electric vehicle capacity, user's expected state of charge, state of charge when the user arrives at the charging station, electric vehicle speed, electric vehicle user's planned parking time, charging completion time, and user's arrival time at the charging station. CS ,L CS ,P ch ,C cos,t Respectively represent the number of charging piles, location of charging station, charging power, and charging electricity price; TN ,ν TN ,g TN Respectively represent the length of the road section, the speed of travel, and the traffic volume. PN ,Q PN ,U PN They represent the active power, reactive power and voltage levels of the distribution network nodes respectively.

[0171] S2.3 Electric Vehicle Load Forecasting Based on GAT-TCN Hybrid Algorithm

[0172] The GAT-TCN hybrid algorithm consists of multiple GATs and TCNs running in parallel. As a neural network architecture designed specifically for processing graph-structured data, GAT introduces a self-attention mechanism that dynamically weighs the importance of each node when aggregating information about adjacent nodes. Compared to traditional graph convolutional networks, which aggregate information using fixed weights, GAT is able to focus on more relevant nodes and accurately capture complex relationships in the graph. This attention mechanism not only enables GAT to process large-scale graph data in parallel, significantly improving computational efficiency, but also allows for greater flexibility in learning spatial relationships between node pairs, eliminating the need to rely on fixed rules.

[0173] After GAT completes the spatial feature learning of the graph features, TCN learns the temporal features of the graph features to obtain the complete spatiotemporal relationship of the electric vehicle charging load data.

[0174] GAT establishes associations between graph data to update the spatial features of different subjects. GAT includes an attention mechanism that selects the optimal weights in the hidden layer to reflect the spatial correlation between multiple subjects in the region. Unlike self-attention, the multi-head attention mechanism enables each attention head to focus on different aspects of the input features, achieving parallel computing and speeding up processing. Therefore, this embodiment introduces a multi-head attention mechanism to convert the graph structure data X = {X1, X2, ..., X N},X N ∈R V Converted into the input of the GAT model. For two adjacent nodes i and j, the attention mechanism describing the importance of node j to node i is as follows:

[0175]

[0176] e ij =Attention(Wx i ,Wx j ) (18)

[0177] Attention(Wx i ,Wx j )=LeakyReLU(w T [Wx i ||Wx j ]) (19)

[0178] Where: Attention(·) represents the attention function; W is the learnable weight parameter; λ i is the first-order neighborhood, including the node itself. ij is the normalized attention coefficient from node to node.

[0179] LeakyReLU(x)=max(0.2x,x), w is a trainable parameter, [Wx i ||Wx j ] represents the concatenation operation, and σ is the activation function.

[0180] The purpose of introducing the multi-head attention mechanism is to calculate the attention coefficients of each neighboring node to the node, and then quantize these coefficients and obtain the final results through network training. Combining the above equations, the updated single node features are obtained. The GAT output vector is represented as X′={X 1′ ,X 2′ ,...,X N′},X N′ ∈R V , and its specific calculation formula is as follows:

[0181]

[0182] TCN is a type of neural network designed for processing time series data. Compared to convolutional neural networks, TCN has stronger temporal causality and a more flexible receptive field. The TCN module consists of three main parts: causal convolution, dilated convolution, and residual convolution.

[0183] Causal convolution strictly follows the time order of data. When considering the data at time t, represented as x t , where t=n*1, y t The prediction of t depends only on the data at time t and the data before that. To illustrate this relationship, the data sequence x0, x1...x t , x t is transformed to predict y0,y1,…y tThe specific formula for this conversion is as follows:

[0184] y0,y1,…y t =f(x0,x1...x t ) (twenty one)

[0185] Traditional causal convolution has a narrow receptive field due to its inherent characteristics, making it difficult to capture long-range dependencies and complex features of time series, limiting model performance. To address this, dilated convolution was introduced. Dilated convolution adjusts sampling and expands the receptive field using a dilation factor d. The formula is as follows.

[0186]

[0187] Where: f(i) represents the i-th convolution coefficient, k represents the size of the convolution kernel, x t-d-i Represents the direction data before time t. When constructing the network, the expansion factor is set to d = 2i, where i = 0, 1, 2, ..., n.

[0188] In TCN, dilated convolution often requires additional network layers, which can easily lead to the vanishing gradient problem. To address this problem, a residual module is constructed, which effectively alleviates this problem through shortcut connections between network layers. The residual connection formula is as follows:

[0189] o=σ(x+F(x)) (23)

[0190] Where: x is the input, F(x) represents the residual mapping to be learned, and o is the output of the layer.

[0191] The output from different channels is represented by o a 、o b and o c , and their sizes are different. The outputs of these different channels are integrated into one through the splicing operation to generate a combined TCN vector h t The calculation formula for this splicing operation is as follows:

[0192]

[0193] Where: It's a splicing symbol.

[0194] S3: Solve the two-layer charging station planning model to obtain the charging station configuration points with the overall optimal solution of the upper-layer objective function and the lower-layer objective function. Configure charging stations in rural areas based on the charging station configuration points with the optimal solution.

[0195] Steps to solve the double-deck charging station planning model:

[0196] 1. Parameter initialization

[0197] Input relevant data, including driving information of electric vehicles (such as cruising range R, driving speed v, etc.); population size N; the number range of charging stations [n min ,n max ]; Maximum number of iterations I max ; Initialize PSO parameters: inertia weight w, cognitive coefficient c1 and social coefficient c2.

[0198] 2. Generate initial particle swarm

[0199] For each particle i (i=1,2,...,N):

[0200] Randomly generate position vector x i , each component represents the node position x of a charging station ij , meeting the site selection conditions and geographical scope constraints.

[0201] Randomly initialize the velocity vector v i , whose size and direction affect the movement trend of particles.

[0202] 3. Calculate fitness value

[0203] For each particle i:

[0204] Calculate the upper-level planning objective function F UPP ; Calculate the lower-level planning objective function F LOW .

[0205] Calculate the fitness value F by integrating the upper and lower layer objective functions i =w1F up +w2F low .

[0206] 4. Update individual optimality and global optimality

[0207] For each particle i:

[0208] If F i <F i,pbest (F i,pbest is the individual historical optimal fitness of particle i), update pbest i =x i and F i,pbest .

[0209] Find the particle with the minimum F in the entire particle swarm i,pbest The particle is recorded as the global optimal gbest and its fitness F i,gbest .

[0210] 5. Update particle position and velocity

[0211] For each particle i:

[0212] Speed ​​update:

[0213] v i (t+1)=w×v i (t)+c1×r1×(pbest i -x i (t))+c2×r2×(gbest-x i (t)), where r1 and r2 are random numbers between 0 and 1.

[0214] Position update: x i (t+1)=x i (t)+v i (t+1).

[0215] 6. Check stop conditions

[0216] If the current number of iterations t=I max , stop iteration; otherwise, t=t+1, return to step 3.

[0217] 7. Output results

[0218] The final rural charging station planning scheme is the particle position corresponding to the global optimal solution gbest, which includes information such as the site selection and capacity configuration of rural charging stations.

[0219] Figure 2 This is a framework diagram for charging station planning based on typical rural scenarios. The proposed method uses a GAT-TCN hybrid algorithm to accurately predict charging load, providing a reliable reference for charging station layout and capacity configuration. A two-tiered rural charging station planning model is constructed, focusing on optimizing costs for builders at the upper level and minimizing charging costs for users at the lower level. Using a particle swarm algorithm, the model is solved to arrive at the optimal rural charging station planning solution.

[0220] To verify the effectiveness of the proposed rural charging station planning scheme, this embodiment conducts a simulation experiment in the IEEE 33-node distribution network system. This embodiment selects three typical rural areas for the experiment, and the selected area is about 12km. 2 , there are 35 road network nodes in total. Rural distribution network topology based on IEEE-33 nodes (such as Figure 3 The area is divided into three villages: sparse village, tourist village and service village (with a highway service area), with a total of 100 electric vehicles.

[0221] In order to verify the prediction accuracy of the GAT-TCN hybrid algorithm model, simulation was performed on the validation set, and the accuracy change curve was obtained as follows: Figure 4 shown.

[0222] Depend on Figure 4As can be seen, the training accuracy curve undergoes two phases: rapid rise and convergence. Initially, as the number of iterations increases, prediction accuracy rapidly increases. During this phase, the model quickly captures and learns the complex interactions and spatial dependencies between multiple sources of information, such as road topology and charging stations, to better predict topological nodes. After approximately 100 iterations, the prediction accuracy curve gradually stabilizes, with an average accuracy of 92.15%. This demonstrates the proposed method's ability to accurately predict topological nodes in complex network environments.

[0223] Figure 5 The following are the charging load demand forecasts for three functional area villages on a specific day. As shown in the figure, the tourist village reaches its peak charging load at approximately 239 kW around 11:00 AM; the service village reaches a peak charging load of approximately 225 kW around 9:00 PM; and the decentralized village's charging load peaks at only approximately 168 kW, with relatively little variation throughout the day. The decentralized village's charging load is lower than that of the tourist and service villages, with a relatively flat charging load curve and a relatively low overall value. Because decentralized villages have a relatively sparse population, fewer electric vehicles are owned, and residents' travel needs and charging behavior are more dispersed, there are no distinct peak charging periods. The tourist village's charging load curve exhibits distinct peaks and valleys. During the day, especially between 8 and 9:00 AM, the charging load rises rapidly. This is due to the increased number of tourists and the increased use of electric vehicles, which requires tourists to charge their vehicles during their visit. Due to the high volume of vehicles passing through the service area, the demand for vehicles to stop for rest, refuel, and charge increases during the day, especially in the afternoon and evening, leading to an increase in charging load.

[0224] In order to clarify the relationship between the total cost of rural charging station planning and the number of charging stations to be built, the corresponding total cost is calculated by changing the number of charging stations to be built. Figure 6 shown. Figure 7 The location results of the optimal rural charging station planning are given.

[0225] From the perspective of cost and location layout, the planning of rural charging stations needs to take into account both the number of construction and location factors. In terms of cost, as the number of charging stations increases from 2 to 8, the total planning cost shows a trend of first decreasing and then increasing. The total cost increases from about 9.5×10 6 The yuan rose to about 10.7×10 6 The construction and operation costs increase with the number of charging stations. Increasing the number of charging stations can reduce user costs. Figure 7As can be seen, CS1 is located in node area 19, a sparsely populated village. Charging demand in this area is relatively dispersed due to the sparseness of villages. This charging station effectively covers the surrounding area, reducing user charging distance costs. CS2 is located in node area 3, a sparsely populated village, meeting the specific charging needs of this area. CS3 is located in node area 8, a tourist village. Considering the potentially large charging demand in tourist villages, this location better serves tourists and their vehicles. CS4 is located in node area 29, a tourist village, providing a wide coverage area to meet the charging load of tourist villages. CS5 is located in node area 14, a service village. Service villages typically have a significant flow of people and vehicles, and this charging station can ensure charging service within the area. Combining cost and location information reveals that when planning rural charging stations, it is necessary to consider the charging demand characteristics of different village types, balance cost variations, and rationally determine the number and location of charging stations. Appropriately increasing the number of charging stations in tourist and service villages with high demand and placing them in key node areas can reduce user costs while, to a certain extent, controlling the growth of construction and operating costs, maximizing the benefits of rural charging station planning.

[0226] Example 2

[0227] like Figure 8 As shown, the charging station planning system based on typical rural scene division in this embodiment includes a model building module, a solution module and an execution module;

[0228] A model construction module is used to construct a two-layer charging station planning model based on an upper-layer objective function and upper-layer constraints, and a lower-layer objective function and lower-layer constraints; the upper-layer objective function aims to minimize the builder's investment cost, and the upper-layer constraints include the power constraint and the charging station footprint constraint; the lower-layer objective function aims to minimize the user's charging cost, and the lower-layer constraints include the remaining power constraint of the electric vehicle and the queue waiting time constraint;

[0229] The solution module is used to determine the target rural type, obtain basic data on the target rural area, and solve the two-tier charging station planning model based on this data. During the solution process, the charging load of the charging station is predicted using a prediction model based on the target rural area data to obtain the overall optimal solution of the upper and lower objective functions. The overall optimal solution of the upper and lower objective functions is used to represent the optimal configuration point planning scheme for charging stations in rural areas. The solution module includes a data acquisition and processing module, a charging load prediction module, a charging station layout optimization module, a constraint management module, and a system visualization module.

[0230] Data Collection and Processing Module: This module collects geographic information about rural areas, power system data, electric vehicle-related data, and user data. The collected data is cleaned, preprocessed, and standardized to remove outliers and erroneous data, converting data from different sources into a unified format and scale.

[0231] The Charging Load Forecasting Module uses GAT to extract spatial features from a coupled network consisting of charging stations, transportation nodes, electric vehicle nodes, and distribution network nodes in rural areas, accurately capturing the complex spatial relationships between nodes. The extracted spatial features are then fed into a Transitional Network (TCN). Using techniques such as causal convolution, dilated convolution, and residual convolution, the TCN learns the temporal characteristics of the data in a time series order, accurately predicting the charging load at each charging station over different time periods.

[0232] Charging station layout optimization module:

[0233] Steps for PSO to solve the rural charging station planning model:

[0234] 1. Parameter initialization

[0235] Input relevant data, including driving information of electric vehicles (such as cruising range R, driving speed v, etc.); population size N; the number range of charging stations [n min ,n max ]; Maximum number of iterations I max ; Initialize PSO parameters: inertia weight w, cognitive coefficient c1 and social coefficient c2.

[0236] 2. Generate initial particle swarm

[0237] For each particle i (i=1,2,...,N):

[0238] Randomly generate position vector x i , each component represents the node position x of a charging station ij , meeting the site selection conditions and geographical scope constraints.

[0239] Randomly initialize the velocity vector v i , whose size and direction affect the movement trend of particles.

[0240] 3. Calculate fitness value

[0241] For each particle i:

[0242] Calculate the upper-level planning objective function F UPP ; Calculate the lower-level planning objective function F LOW .

[0243] Calculate the fitness value F by integrating the upper and lower layer objective functions i =w1F up+w2F low .

[0244] 4. Update individual optimality and global optimality

[0245] For each particle i:

[0246] If F i <F i,pbest (F i,pbest is the individual historical optimal fitness of particle i), update pbest i =x i and F i,pbest .

[0247] Find the particle with the minimum F in the entire particle swarm i,pbest The particle is recorded as the global optimal gbest and its fitness F i,gbest .

[0248] 5. Update particle position and velocity

[0249] For each particle i:

[0250] Speed ​​update:

[0251] v i (t+1)=w×v i (t)+c1×r1×(pbest i -x i (t))+c2×r2×(gbest-x i (t)), where r1 and r2 are random numbers between 0 and 1.

[0252] Position update: x i (t+1)=x i (t)+v i (t+1).

[0253] 6. Check stop conditions

[0254] If the current number of iterations t=I max , stop iteration; otherwise, t=t+1, return to step 3.

[0255] 7. Output results

[0256] The final rural charging station planning scheme is the particle position corresponding to the global optimal solution gbest, which includes information such as the location and capacity configuration of rural charging stations.

[0257] Constraint Management Module: Manages and maintains various constraints involved in the charging station planning process, including equational constraints such as the power flow equation of the power system. Through accurate power system models, it ensures the power balance and stable operation of the power system during the operation of the charging station.

[0258] System Visualization Module: This module displays charging load forecast curves, cost analysis charts, and user satisfaction indicators through various charts, providing a quantitative basis for scheme evaluation. It also offers interactive features, allowing users to easily view detailed information and data about charging stations. It also allows for dynamic demonstrations, showing how the construction and operation of charging stations change over time under different planning schemes.

[0259] The execution module is used to configure charging stations in rural areas according to the charging station configuration points of the optimal solution.

[0260] Example 3

[0261] An embodiment of the present invention discloses a device for determining charging station planning based on the division of typical rural scenarios. The device may include a processor and a memory, wherein the memory stores computer instructions, and the processor is used to execute the computer instructions stored in the memory. When the computer instructions are executed by the processor, the electronic device implements the steps of the method described in the above embodiment and can achieve technical effects consistent with the above method.

[0262] The memory may include computer system readable media in the form of volatile memory, such as random access memory (RAM) and / or cache memory. The device may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the memory may be used to read and write non-removable, non-volatile magnetic media (commonly referred to as a "hard drive"). A program / utility having a set (at least one) of program modules may be stored in, for example, the memory, such program modules including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. The program modules typically perform the functions and / or methods of the embodiments described herein.

[0263] The processor executes various functional applications and data processing by running the programs stored in the memory, such as implementing the method provided in the first embodiment of the present invention.

[0264] Example 4

[0265] Embodiment 4 of the present invention further provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method described in the above embodiment are implemented and the technical effect consistent with the above method can be achieved.

[0266] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or components, or any combination thereof. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by an instruction execution system, device or device or used in combination with it.

[0267] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0268] Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0269] The computer program code for performing the operations of the present invention can be written in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0270] Of course, the computer-executable instructions of a storage medium provided by an embodiment of the present invention are not limited to the above method operations, but can also execute related operations in the method provided by any embodiment of the present invention.

[0271] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A charging station planning method based on typical rural scene division, characterized by: The following steps are involved: A two-layer charging station planning model is constructed based on an upper-layer objective function and upper-layer constraints, as well as a lower-layer objective function and lower-layer constraints. The upper-layer objective function aims to minimize the builder's investment cost, and the upper-layer constraints include the power constraint and the area constraint of the charging station. The lower-layer objective function aims to minimize the user's charging cost, and the lower-layer constraints include the remaining power constraint of the electric vehicle and the waiting time constraint in the queue. Determine the target rural type, obtain basic data of the target rural area, and solve the two-layer charging station planning model based on the basic data of the target rural area. During the solution process, the charging load of the charging station is predicted through a prediction model based on the basic data of the target rural area to obtain the overall optimal solution of the upper and lower objective functions; the overall optimal solution of the upper and lower objective functions is used to represent the optimal configuration point planning scheme for configuring charging stations in rural areas.

2. The charging station planning method according to claim 1, characterized in that: The upper objective function is: min f upp =C 1,θ +C 2,θ Among them, C 1,θ represents the construction cost; C 2,θ represents operating costs; Construction cost C 1,θ and operating costs C 2,θ The calculation formula is: in, represents the area occupied by charging station i, Represents the charging station capacity of charging station i, i=1,2,3....N CS , N CS is the number of charging stations, a θ represents the unit land price at the construction location of charging station i under the rural type θ, b θ represents the unit cost of the charging pile and related supporting equipment of the construction unit under the rural type, c θ represents the cost of distribution transformers and cables under the rural type; k1 and k2 represent cost conversion coefficients, μ represents the equipment maintenance rate; y represents the operating life of the charging station; The upper-level constraints include: Power constraints of charging stations: Among them, P i,j represents the output power of the jth charging device in charging station i, represents the number of charging devices in charging station i; P i,max,θ is the maximum output power of charging station i under the rural type of θ; Charging station floor space constraints: Among them, S i,j is the area occupied by the jth charging equipment in charging station i; S i,max is the maximum floor space of charging station i.

3. The charging station planning method according to claim 2, characterized in that: The lower layer objective function is: minf low =C EV Among them, C EV Charging costs for electric vehicle users; Charging cost C for electric vehicle users EV The calculation formula is: Among them, C u,t The cost of electric energy lost when driving an electric vehicle to a charging station; The equivalent time loss cost required for electric vehicle users to travel to charging stations; is the necessary queuing time cost faced by electric vehicle users after driving to the target charging station; M represents the number of all electric vehicle users in the planning area; Electricity loss cost C u,t The calculation formula is: Among them, D uj is the distance between electric vehicle user u and the best charging station; L u is the total cruising range of the electric vehicle u; C u is the battery capacity of the electric vehicle u; η is the vehicle charging and discharging efficiency coefficient; C cos is the unit electric energy loss cost of electric vehicle u; Equivalent time loss cost The calculation formula is: Among them, λ tc is the loss cost per unit time; V u is the speed of the electric vehicle; Necessary queuing time cost The calculation formula is: Among them, T u Waiting time for charging: The lower level constraints include: Remaining power constraints of electric vehicles: S u ×L u ≥D uj Among them, S u The remaining power of the electric vehicle; Queue waiting time constraints: Among them, T u Waiting time for electric vehicle users, The maximum waiting time in queue for electric vehicle users.

4. The charging station planning method according to claim 1, characterized in that: The target rural types include tourist villages, sparsely populated villages, and villages with service areas; the judging indicators are: the presence of service areas, the tourism heat index K1, and the residential dispersion index K2; If a rural area has no service area and the tourism heat index K1 exceeds a specific threshold, it is determined to be a tourist village; if a rural area has no service area and the tourism heat index does not exceed the specific threshold K1, but the residential dispersion index K2 exceeds a specific threshold, it is determined to be a sparse village; When a rural area has a service area, it is determined to be a village with a service area.

5. The charging station planning method according to claim 1, characterized in that: The prediction model adopts an electric vehicle load prediction model based on the GAT-TCN hybrid algorithm. First, the rural coupling network is represented by a graph structure. The spatial features of the graph are learned through the GAT model. The extracted spatial features are input into the TCN model. The TCN model learns the temporal features of the graph features and predicts the charging load of each charging station in different time periods. The graph feature is: X=[X EV ,X CS ,X TN ,X PN ] Among them, X EV represents the electric vehicle node characteristics, X CS represents the charging station node characteristics, X TN Traffic node characteristics, X PN Distribution network node characteristics, t represents time, l u Indicates the position of the electric vehicle, S u Indicates the SOC of electric vehicle battery, C u Indicates the battery capacity of electric vehicles, Indicates the user's desired state of charge. Indicates the state of charge when the user arrives at the charging station. represents the speed of the electric vehicle, represents the planned parking time of electric vehicle users, Indicates charging completion time. Indicates the time when the user arrives at the charging station; s CS Indicates the number of charging piles in the charging station, L CS Indicates the location of the charging station, Indicates charging power, C cos,t Indicates the charging electricity price; l TN represents the length of the road segment, ν TN Indicates the travel speed, g TN represents the traffic volume; P PN Indicates the active power of the distribution network node, Q PN Represents the reactive power of the distribution network node, U PN Indicates the voltage at the distribution network node.

6. The charging station planning method according to claim 1, characterized in that: The specific steps of solving the double-deck charging station planning model include: Initialize the parameters of the double-deck charging station planning model; Generate an initial particle swarm, where each particle represents the location of a charging station; Calculate the value of the upper layer objective function and the value of the lower layer objective function for each particle, and take the weighted sum of the value of the upper layer objective function and the value of the lower layer objective function as the fitness value; find the particle with the best fitness value in the entire particle swarm; The particles are iteratively updated, and the particle with the best fitness value is output at the end of the iteration to obtain the optimal solution.

7. A charging station planning system based on typical rural scene division, characterized by: include: A model construction module is used to construct a two-layer charging station planning model based on an upper-layer objective function and upper-layer constraints, and a lower-layer objective function and lower-layer constraints; the upper-layer objective function aims to minimize the builder's investment cost, and the upper-layer constraints include the power constraint and the charging station footprint constraint; the lower-layer objective function aims to minimize the user's charging cost, and the lower-layer constraints include the remaining power constraint of the electric vehicle and the queue waiting time constraint; The solution module is used to determine the target rural type, obtain the basic data of the target rural area, and solve the two-layer charging station planning model based on the basic data of the target rural area. During the solution process, the charging load of the charging station is predicted through the prediction model based on the basic data of the target rural area to obtain the overall optimal solution of the upper and lower objective functions; the overall optimal solution of the upper and lower objective functions is used to represent the optimal configuration point planning scheme for configuring charging stations in rural areas.

8. The charging station planning system according to claim 7, characterized in that: The upper objective function is: min f upp =C 1,θ +C 2,θ Among them, C 1,θ represents the construction cost; C 2,θ represents operating costs; Construction cost C 1,θ and operating costs C 2,θ The calculation formula is: in, represents the area occupied by charging station i, Represents the charging station capacity of charging station i, i=1,2,3....N CS , N CS is the number of charging stations, a θ represents the unit land price at the construction location of charging station i under the rural type θ, b θ represents the unit cost of the charging pile and related supporting equipment of the construction unit under the rural type, c θ represents the cost of distribution transformers and cables under the rural type; k1 and k2 represent cost conversion coefficients, μ represents the equipment maintenance rate; y represents the operating life of the charging station The upper-level constraints include: Power constraints of charging stations: Among them, P i,j represents the output power of the jth charging device in charging station i, represents the number of charging devices in charging station i; P i,max,θ is the maximum output power of charging station i under the rural type of θ; Charging station floor space constraints: Among them, S i,j is the area occupied by the jth charging equipment in charging station i; S i,max is the maximum floor space of charging station i.

9. The charging station planning system according to claim 8, characterized in that: The lower layer objective function is: minf low =C EV Among them, C EV Charging costs for electric vehicle users; Charging cost C for electric vehicle users EV The calculation formula is: Among them, C u,t The cost of electric energy lost when driving an electric vehicle to a charging station; The equivalent time loss cost required for electric vehicle users to travel to charging stations; is the necessary queuing time cost faced by electric vehicle users after driving to the target charging station; M represents the number of all electric vehicle users in the planning area; Electricity loss cost C u,t The calculation formula is: Among them, D uj is the distance between electric vehicle user u and the best charging station; L u is the total cruising range of the electric vehicle u; C u is the battery capacity of the electric vehicle u; η is the vehicle charging and discharging efficiency coefficient; C cos is the unit electric energy loss cost of electric vehicle u; Equivalent time loss cost The calculation formula is: Among them, λ tc is the loss cost per unit time; V u is the speed of the electric vehicle; Necessary queuing time cost The calculation formula is: Among them, T u Waiting time for charging: The lower level constraints include: Remaining power constraints of electric vehicles: S u ×L u ≥D uj Among them, S u The remaining power of the electric vehicle; Queue waiting time constraints: Among them, T u Waiting time for electric vehicle users, The maximum waiting time in queue for electric vehicle users.

10. The charging station planning system according to claim 7, characterized in that: The target rural types include tourist villages, sparsely populated villages, and villages with service areas; the judging indicators are: the presence of service areas, the tourism heat index K1, and the residential dispersion index K2; If a rural area has no service area and the tourism heat index K1 exceeds a specific threshold, it is determined to be a tourist village; if a rural area has no service area and the tourism heat index does not exceed the specific threshold K1, but the residential dispersion index K2 exceeds a specific threshold, it is determined to be a sparse village; When a rural area has a service area, it is determined to be a village with a service area.

11. The charging station planning system according to claim 7, characterized in that: The prediction model adopts an electric vehicle load prediction model based on the GAT-TCN hybrid algorithm. First, the rural coupling network is represented by a graph structure. The spatial features of the graph are learned through the GAT model. The extracted spatial features are input into the TCN model. The TCN model learns the temporal features of the graph features and predicts the charging load of each charging station in different time periods. The graph feature is: X=[X EV ,X CS ,X TN ,X PN ] Among them, X EV represents the electric vehicle node characteristics, X CS represents the charging station node characteristics, X TN Traffic node characteristics, X PN Distribution network node characteristics, t represents time, l u Indicates the position of the electric vehicle, S u Indicates the SOC of electric vehicle battery, C u Indicates the battery capacity of electric vehicles, Indicates the user's desired state of charge. Indicates the state of charge when the user arrives at the charging station. represents the speed of the electric vehicle, represents the planned parking time of electric vehicle users, Indicates charging completion time. Indicates the time when the user arrives at the charging station; s CS Indicates the number of charging piles in the charging station, L CS Indicates the location of the charging station, Indicates the charging power, C cos,t Indicates the charging electricity price; l TN represents the length of the road segment, ν TN Indicates the travel speed, g TN represents the traffic volume; P PN Indicates the active power of the distribution network node, Q PN Represents the reactive power of the distribution network node, U PN Indicates the voltage at the distribution network node.

12. The charging station planning system according to claim 7, characterized in that: The specific steps of solving the double-deck charging station planning model include: Initialize the parameters of the double-deck charging station planning model; Generate an initial particle swarm, where each particle represents the location of a charging station; Calculate the value of the upper layer objective function and the value of the lower layer objective function for each particle, and take the weighted sum of the value of the upper layer objective function and the value of the lower layer objective function as the fitness value; find the particle with the best fitness value in the entire particle swarm; The particles are iteratively updated, and the particle with the best fitness value is output at the end of the iteration to obtain the optimal solution.

13. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

14. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images