A charging facility planning method, device and medium based on machine learning

Abstract

The present invention relates to the technical field of urban planning, and specifically to a charging facility planning method, device, and medium based on machine learning. The method comprises the following steps: gridding the planning area, including multiple charging service grids; predicting the number of vehicles in use and the demand for charging facilities to obtain charging demand; using a clustering algorithm to obtain multiple charging blocks with balanced charging demand; constructing an objective function based on maximizing the revenue of each charging block, constructing constraints based on full coverage of charging demand, calling the Aurora optimization algorithm to obtain an optimal planning scheme for charging facilities, and completing charging facility planning using the optimal planning scheme for charging facilities. The method of the present invention can solve the problems of incomplete analysis and unreasonable optimization zoning in the prior art, comprehensively consider various factors affecting the layout of charging facilities, thereby achieving efficient layout and resource allocation of the charging network, and significantly improving the overall quality of charging services and user satisfaction.

Description

Technical Field

[0001] The present invention relates to the field of urban planning technology, and in particular to a charging facility planning method, device, and medium based on machine learning. Background Art

[0002] With rising global environmental awareness and adjustments to energy mix, electric vehicles (EVs) are rapidly gaining popularity worldwide due to their zero-emission and low-noise characteristics. However, the development of charging infrastructure faces numerous challenges, particularly in the absence of unified urban planning. The spatial distribution of charging facilities often exhibits uneven distribution, directly impacting the quality and efficiency of charging services.

[0003] Existing planning and design schemes are all based on demand, with site determination, capacity selection, and equipment configuration. However, since charging stations cannot be located independently, the optimal site areas recommended by the plan often lack suitable parking resources or face problems such as insufficient grid support capacity.

[0004] In summary, there is an urgent need for a charging facility planning method, equipment, and medium based on machine learning to solve existing problems. Summary of the Invention

[0005] The present invention aims to provide a charging facility planning method, device, and medium based on machine learning. The specific technical solutions are as follows:

[0006] A charging facility planning method based on machine learning includes the following steps:

[0007] S100: Divide the planning area into grids, including multiple charging service grids with the same service range;

[0008] S200: Carry out vehicle ownership prediction and charging facility demand prediction in sequence based on the charging service grid to obtain charging demand of the charging service grid;

[0009] S300: A clustering algorithm is used to perform a fusion analysis of the grids in the planning area. The charging resource conditions and service capabilities of the planning area are comprehensively considered. Based on the charging demand and spatial distribution characteristics of the charging service grid, multiple charging blocks with balanced charging demand are clustered.

[0010] S400: Construct an objective function based on maximizing the revenue of stations in each charging block area, construct constraints based on full coverage of charging demand and optimal utilization of charging resources, call the Aurora optimization algorithm to obtain the optimal planning scheme for charging facilities, and use the optimal planning scheme for charging facilities to complete charging facility planning.

[0011] Optionally, in S100 , the charging service grid is a square grid of uniform size, with a side length ranging from 0.5 km to 1 km.

[0012] Optionally, in S200, the vehicle ownership prediction includes sequentially performing a car ownership prediction and an electric vehicle ownership prediction;

[0013] S201, Car Ownership Forecast: For private cars, the per capita disposable income level in the planning area is predicted by the growth rate method in the target year. The benchmark area is determined based on the per capita disposable income level, and the data is verified using the population age structure and the average number of cars per household. For other types of cars, the growth rate method is used for forecasting, and the data is adjusted based on the actual development of different types of cars.

[0014] S202, Electric Vehicle Ownership Forecast: Based on the vehicle ownership forecast results, the Bass diffusion model is used to carry out electric vehicle ownership forecast. The cumulative curve expression and density expression of the Bass diffusion model are as follows:

[0015] ;

[0016] ;

[0017] in, represents the cumulative number of electric vehicles adopted before time t; Indicates the maximum market potential; represents the innovation coefficient; represents the imitation coefficient; represents the number of electric vehicle adopters at time t.

[0018] Optionally, in S200 , the charging facility demand forecast includes a total charging facility demand forecast and a charging facility demand forecast for each grid;

[0019] S203, Forecasting the Total Demand for Charging Facilities: Different vehicle-to-charging pile ratios are assigned to different types of vehicles. Based on the number of electric vehicles in use and a differentiated approach, the annual charging demand for each type of electric vehicle is estimated. The total charging power of the charging facilities in the planned area is calculated to complete the forecast of the total demand for charging facilities.

[0020] S204, forecasting demand for charging facilities in each grid: Based on the actual annual power demand, target demand allocation is made for the total number of charging facilities according to the population, road network density, and refined oil consumption of each grid. The forecast formula for charging facility demand in each grid is as follows:

[0021] The actual charging demand ratio of a certain grid = the population ratio of a certain grid × 0.4 + the refined oil consumption ratio of a certain grid + 0.4 + the road network density ratio of a certain grid × 0.2.

[0022] Optionally, in S300, the process of using a clustering algorithm to perform fusion analysis on the grids of the planning area is as follows:

[0023] S301, initialization: Referring to the level of refined oil consumption at gas stations in the region, combined with the forecast results of electric vehicle ownership, matching the charging service capacity of the charging station, and calculating the target year single-station service vehicle, electrification rate, and charging service demand density value. The expression is as follows:

[0024] ;

[0025] in, Indicates the target annual number of vehicles served by a single charging station in the region. Indicates the electrification rate of vehicles served by a single charging station in the target year, represents the charging service demand density value, It represents the average number of vehicles served per gas station per year. represents the annual average electrification rate of vehicles served per gas station. Indicates the service range of a single charging station;

[0026] According to the grid division principle, sample grids, the position of the i-th sample grid is , the charging demand is ;pass The most reasonable number k of charging blocks that needs to be further optimized is obtained, and the expression is as follows:

[0027] ;

[0028] in, is the total charging demand in the planning area, and the optimal charging demand for each charging block for:

[0029] ;

[0030] Randomly select k sample grids as the preset initial center points of each charging block area, and the geometric center point of the jth initial center point is used as its coordinate position ;

[0031] S302, Calculate distance and assign grid: distance metric It is a function of spatial distance, which represents the distance between the i-th grid and the j-th center point. The expression is as follows:

[0032] ;

[0033] According to the distance calculation result, the sample grid is assigned to the central grid block with the smallest distance. The expression is as follows:

[0034] ;

[0035] in, represents the charging block index to which the sample grid i is assigned, and arg represents the action corresponding to the optimal strategy;

[0036] S303, update the charging demand of each charging block area: As the sample grid is allocated, the charging demand of each charging block area j is updated. The expression is as follows:

[0037] ;

[0038] in, is the set of sample grids in the jth charging block, For Grid Charging demand;

[0039] S304, update the center position of the charging block area: the center position of the current charging block area is determined by the weighted average position of all small grids in the charging block area, and the update formula is:

[0040] ;

[0041] S305, demand balance: calculate demand error: demand error for each block j Defined as: ;

[0042] Introduce a balancing strategy to ensure that the charging demand of each charging block is close to the target value ;

[0043] Balancing strategy: Adjust the membership of charging blocks based on demand differences, or shift charging service grids from charging blocks with excessive demand to charging blocks with insufficient demand, until the demand deviation reaches an acceptable range.

[0044] Convergence condition: When the demand error of all charging blocks Less than the predetermined threshold When , the algorithm stops iterating, and the expression of the convergence condition is as follows:

[0045] ;

[0046] in, Indicates the maximum allowable value of demand quantity difference;

[0047] S306, output result: When the total demand of each charging block area is close to the target demand amount P, clustering is completed, and k charging blocks with balanced charging demand amounts are output.

[0048] Optionally, in S400 , the objective function is to maximize the annual operating income of the charging station in the target charging block, including charging service fees, V2G discharge response service income, annualized construction cost of the charging station, and annual operating cost. The objective function is expressed as follows:

[0049] ;

[0050] in, is the annualized total return, For charging station operation service income, To convert the construction cost to annual cost, The annual operation and maintenance cost of the charging station;

[0051] The charging station operation service income , the calculation process is as follows:

[0052] ;

[0053] ;

[0054] ;

[0055] in, Indicates the annual charging service fee of the charging station; represents the service revenue of V2G discharge response in one year; The unit price of the service per kilowatt of charging power; represents the service probability coefficient of the station; represents the number of V2G discharges in the region in a year; Indicates the amount of electricity discharged in a single time; represents the service fee for V2G discharge;

[0056] The construction cost of the charging station , the calculation process is as follows:

[0057] ;

[0058] ;

[0059] ;

[0060] ;

[0061] in, represents the average discount rate; Indicates the depreciation period; Indicates the charging stations; Indicates the number of alternative charging stations within the charging block area; represents the infrastructure operation cost; represents the purchase cost of charging piles at the i-th charging station; represents the cost of supporting facilities; The unit price of the land occupied by the station; Indicates the ratio coefficient between the charging power and the floor area of the station; It represents the comprehensive cost per watt of charging infrastructure construction; is the equivalent investment coefficient of supporting facilities cost;

[0062] The operation and maintenance costs of the charging station , the calculation process is as follows:

[0063] ;

[0064] ;

[0065] in, is the annual unit labor cost; is a step function, Indicates the minimum number of fast charging piles at the station, rounded down; Charging infrastructure construction and annual maintenance cost per kilowatt.

[0066] Optionally, in S400 , the constraints include a power supply and demand balance constraint, a charging station power constraint, and a charging station distance constraint;

[0067] The expression of power supply and demand balance constraint is as follows:

[0068] ;

[0069] in, It is The number of alternative charging station sites within the block charging area; For the Block charging: The total amount of charging demand in the block area;

[0070] The expression of the charging station power constraint is as follows:

[0071] ;

[0072] in, Indicates that the total power of the charging piles that can be connected to the station should not exceed the maximum power allowed by the power grid;

[0073] The expression of the charging station distance constraint is as follows:

[0074] ;

[0075] Among them, when hour, , indicating that the site is not selected, that is, the service range impact of the site does not need to be considered; when hour, , indicating that the site is selected, and the distance between stations must not be less than the minimum value .

[0076] Optionally, in S400 , the Aurora optimization algorithm is called to obtain the optimal planning scheme for charging facilities. The process is as follows:

[0077] S401, Initialization: Randomly initialize the charging station power allocation value Initialize the auroral group position and velocity, where each position represents a power allocation scheme and the velocity represents the update rate;

[0078] S402, calculate the objective function: For each charging station power allocation scheme, calculate the objective function:

[0079] ;

[0080] in, Indicates the Total revenue of the station in each block area;

[0081] S403, Aurora group position update: Update the power allocation of each charging station:

[0082] ;

[0083] in, is the optimal location of the current charging station i; For the optimal power allocation solution; and are adjustment factors, used to control the convergence speed;

[0084] S403, check the constraints: after each update, check whether the power allocation meets the constraints. If not, a penalty function is used to impose a cost penalty on the power allocation solution that does not meet the constraints. The penalty cost is The expression is as follows:

[0085] ;

[0086] in: is the penalty factor, which controls the intensity of the penalty; when When , the penalty function value is zero, indicating that the constraint condition is satisfied; otherwise, the penalty cost increases, which makes the benefit value of the objective function smaller and reduces the excellence of the power allocation scheme;

[0087] S404, termination of iteration: When the maximum number of iterations is reached or the objective function converges to a predetermined accuracy, the algorithm is stopped to obtain the optimal power allocation solution for each charging station. .

[0088] Additionally, the present invention also includes a computer device comprising a memory and a processor;

[0089] The memory is used to store a computer program that can be executed on the processor;

[0090] The processor is used to implement the steps of the charging facility planning method as described above when executing the computer program.

[0091] In addition, the present invention also includes a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the charging facility planning method as described above are implemented.

[0092] The application of the technical solution of the present invention has the following beneficial effects:

[0093] (1) The present invention discloses a charging facility planning method based on machine learning. The method uses a clustering algorithm to optimize the service radius and service range of charging stations. This method can effectively solve the problems of incomplete analysis and unreasonable optimization zoning in the existing technology, and comprehensively consider various factors affecting the layout of charging facilities, thereby achieving efficient layout and resource allocation of the charging network, and significantly improving the overall quality of charging services and user satisfaction.

[0094] (2) The method of the present invention aims to solve the problem that the current prediction of the number of private cars of the incremental subject is not accurate enough. It studies the logical relationship between per capita disposable income and the number of private cars per thousand people, and proposes to scientifically predict the per capita disposable income level of the study area in the target year by the growth rate method, so as to determine the level of the benchmark area based on the per capita disposable income value. At the same time, the data is verified by methods such as population age structure and the average number of cars per household, so that the prediction data of the incremental subject of private cars in the target year are relatively reliable and reasonable.

[0095] (3) The method of the present invention addresses the problem of unclear matching between charging demand and available charging resources in current planning. Taking into account the differences in driving characteristics and site conditions of different types of electric vehicles, a differentiated supply and demand balance matching method for charging facilities is proposed. Electric vehicles are classified and the charging facilities of different types of vehicles are matched in different ways. At the same time, the interconnection and interoperability of the supply capacity of public charging facilities and dedicated charging facilities are taken into account to maximize the utilization of public charging facilities.

[0096] (4) The method of the present invention uses the k-means clustering algorithm to perform grid division on the study area, dividing the area into several blocks with relatively balanced charging demand and resource conditions. The division method provided by the present invention comprehensively considers factors such as charging demand, parking space, power supply capacity, road network resources, and the service capacity of existing charging facilities in each charging block, providing a scientific and reasonable regional division basis for the subsequent planning of charging station sites. In this way, human intervention in the regional division process can be effectively reduced, ensuring the objectivity and accuracy of the regional division, thereby providing scientific and powerful support for the site selection of charging facilities.

[0097] (5) The method of the present invention designs a planning scheme based on the Aurora optimization algorithm to optimally plan the charging station sites in each block area. The method of the present invention takes the maximization of the total revenue of the stations in each block area as the objective function and ensures full coverage of charging demand as a constraint condition. On this basis, combined with the specific resource conditions of each block area, the site selection and capacity optimization configuration of the charging station are efficiently completed to ensure that while meeting the charging demand, the economy and construction convenience of the charging station are improved. The method of the present invention can solve the problem that the traditional charging station site selection only starts from the single positive thinking of charging demand and cannot achieve the optimal site selection and optimal configuration on the resource side due to the site resource endowment.

[0098] In addition to the above-described objects, features and advantages, the present invention has other objects, features and advantages. The present invention will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0100] Figure 1 It is a flowchart of the steps of the charging facility planning method in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0101] In order to enable those skilled in the art to better understand the present invention, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0102] like Figure 1 As shown, this embodiment provides a charging facility planning method based on machine learning, including the following steps:

[0103] S100: Divide the planning area into grids, including multiple charging service grids with the same service range;

[0104] S200: Carry out vehicle ownership prediction and charging facility demand prediction in sequence based on the charging service grid to obtain charging demand of the charging service grid;

[0105] S300: A clustering algorithm is used to perform a fusion analysis of the grids in the planning area. The charging resource conditions and service capabilities of the planning area are comprehensively considered. Based on the charging demand and spatial distribution characteristics of the charging service grid, multiple charging blocks with balanced charging demand are clustered.

[0106] S400: Construct an objective function based on maximizing the revenue of stations in each charging block area, construct constraints based on full coverage of charging demand and optimal utilization of charging resources, call the Aurora optimization algorithm to obtain the optimal planning scheme for charging facilities, and use the optimal planning scheme for charging facilities to complete charging facility planning.

[0107] Specifically, in S100, the charging service grid is a square grid of uniform size, with a side length ranging from 0.5km to 1km. In this embodiment, when collecting samples, the study area is divided into grids, which can unify all characteristic variables into the same grid, ensure data accuracy, and bring certain conveniences to data collection. At the same time, the size of the grid is particularly important for predictive analysis. If data analysis is performed based on communities or cells, the differences between units are small, and if the grid is too large, the accuracy of the sample space will be affected.

[0108] Furthermore, in this embodiment, the preferred service range of the charging service grid includes a city-level area and a county-level area. The city-level area has a range of 1 km×1 km, and the county-level area has a range of 0.5 km×0.5 km.

[0109] Specifically, in S200 , the vehicle ownership prediction includes sequentially performing the car ownership prediction and the electric vehicle ownership prediction;

[0110] S201, Car Ownership Forecast: For private cars, the per capita disposable income level in the planning area is predicted by the growth rate method in the target year. The benchmark area is determined based on the per capita disposable income level, and the data is verified using the population age structure and the average number of cars per household. For other types of cars, the growth rate method is used for forecasting, and the data is adjusted based on the actual development of different types of vehicles (for example, urban buses and cruising taxis have basically reached saturation).

[0111] S202, Electric Vehicle Ownership Forecast: Based on the vehicle ownership forecast results, the Bass Diffusion Model is used to carry out electric vehicle ownership forecast.

[0112] The Bass Diffusion Model is a mathematical model used to predict the spread and adoption of new products or technologies in the market. After years of research and refinement, it has become a valuable tool in fields such as communication, marketing, and management. The model is based on the following assumptions: product adopters can be divided into two categories: "innovators," who are willing to try new products and set trends in the market; and "followers," who only consider adopting a new product after seeing others use it. The number of these two types of adopters is fixed, and the total market size is known.

[0113] In this embodiment, the cumulative curve expression and density expression of the Bass diffusion model are as follows:

[0114] ;

[0115] ;

[0116] in, represents the cumulative number of electric vehicles adopted before time t; Indicates the maximum market potential; represents the innovation coefficient; represents the imitation coefficient; represents the number of electric vehicle adopters at time t.

[0117] It should be noted that this embodiment takes into account the need to predict electric vehicle sales based on historical data while introducing the impact of charging infrastructure on electric vehicle sales. Therefore, the Bass diffusion model is used to predict electric vehicle sales. The Bass diffusion model is simple in method and accurate in results. The generalized Bass diffusion model can quantitatively describe the impact of external factors.

[0118] Specifically, in S200 , the charging facility demand forecast includes the total charging facility demand forecast and the charging facility demand forecast for each grid;

[0119] S203, forecast of total demand for charging facilities: Different vehicle-to-pile ratios are divided according to different types of vehicles. Based on the number of electric vehicles and a differentiation method (the differentiation method in this embodiment includes the mileage charging demand method and the vehicle-to-pile ratio method), the annual charging demand for each type of electric vehicle is estimated. The total charging power of the charging facilities in the planned area is calculated to complete the forecast of total demand for charging facilities.

[0120] It should be noted that different types of electric vehicles have different charging requirements due to their different driving characteristics and site conditions, and differentiated charging facility matching solutions are needed to meet them. The matching solutions for charging facilities and vehicle models in the plan are as follows:

[0121] 1) Buses, passenger vehicles, and medium and heavy-duty trucks generally have fixed parking lots. Due to their size, they are generally difficult to charge at public charging stations. Special charging piles built in parking lots are the main charging method. Other types of vehicles also use special charging stations built in parking lots as a charging method due to their special functions and uses.

[0122] 2) The average daily parking time of private cars is much longer than the driving time, and the average daily mileage is less than 50 kilometers. Users can replenish electricity through private charging piles combined with public charging facilities.

[0123] 3) Taxis, government vehicles, micro trucks, and light trucks have high mobility and use public charging facilities as a charging method.

[0124] This embodiment analyzes influencing factors such as the driving characteristics, parking patterns, driving range, and facility locations of various types of vehicles, and refers to the vehicle configuration principles of some provinces and cities to determine the standard principles for the configuration of various types of charging facilities:

[0125] 1) The routes and mileage of buses and passenger vehicles are relatively fixed. The average daily mileage of buses and passenger vehicles in large and medium-sized cities is 200-250 kilometers. The stops are parking lots at the starting and ending stations. They mainly replenish energy through bus charging stations. Based on factors such as charging time and vehicle idle time, the bus charging station ratio should be no less than 3:1.

[0126] 2) Medium and heavy-duty trucks are generally difficult to charge at public charging stations due to their size and require dedicated charging piles. However, unlike buses and passenger vehicles, their charging time is more flexible. Based on factors such as charging time and vehicle idle time, the ratio of vehicle to charging pile should be no less than 4:1. Other methods of electricity replenishment such as battery swapping are encouraged according to actual conditions.

[0127] 3) For other types of vehicles, such as sanitation vehicles, the ratio of vehicles to charging piles should be no less than 5:1, depending on factors such as charging time and vehicle idle time.

[0128] 4) Private charging facilities shall be constructed with a vehicle-to-charging pile ratio of 4:1.

[0129] In this embodiment, the charging demand of a certain type of electric vehicle in one year is estimated. , the calculation method is as follows:

[0130] ;

[0131] in, Indicates the number of electric vehicles of this type. Indicates the average power consumption per 100 kilometers of this type of electric vehicle. Indicates the average annual mileage of electric vehicles of this type (which can be calculated based on historical data).

[0132] In this embodiment, the average annual mileage of electric private cars and official vehicles is about 10,000 kilometers, and the average power consumption is about 20 kWh / 100 kilometers; the average annual mileage of electric taxis is about 100,000 kilometers, and the average power consumption is about 18 kWh / 100 kilometers; the average annual mileage of micro passenger vehicles and light trucks is about 40,000 kilometers, and the average power consumption is about 20 kWh / 100 kilometers.

[0133] Furthermore, the total charging power of public charging facilities The expression is as follows:

[0134] ;

[0135] in, represents the average annual charging capacity of a certain type of electric vehicle, It indicates the proportion of public sector charging volume of a certain type of electric vehicle to the annual average charging volume, Indicates the daily utilization hours of public charging facilities.

[0136] In this embodiment, the proportion of public-sector charging for private cars to the annual total charging volume is taken as 80%, the proportion of public-sector charging for taxis, official vehicles, micro-passenger vehicles, and light-duty trucks to the annual total charging volume is taken as 100%, and the power utilization rate of public charging infrastructure is considered to be 7.5%.

[0137] S204, Forecast of Charging Facility Demand for Each Grid: Based on the forecast of total charging facility demand, the charging demand of residents who own electric vehicles but do not have private parking spaces or are not equipped with private charging infrastructure within the community is taken into account. The target annual actual power demand is deducted from the existing charging power and the planned AC slow charging pile power. Based on the actual annual power demand, the total amount of charging facilities is allocated according to the population, road network density, and refined oil consumption of each grid. The forecast formula for charging facility demand for each grid is as follows:

[0138] The actual charging demand ratio of a certain grid = the population ratio of a certain grid × 0.4 + the refined oil consumption ratio of a certain grid + 0.4 + the road network density ratio of a certain grid × 0.2.

[0139] Specifically, in S300, the process of using the clustering algorithm to perform fusion analysis on the grids of the planning area is as follows:

[0140] S301, initialization: Referring to the level of refined oil consumption at gas stations in the region, combined with the forecast results of electric vehicle ownership, matching the charging service capacity of the charging station, and calculating the target year single-station service vehicle, electrification rate, and charging service demand density value. The expression is as follows:

[0141] ;

[0142] in, Indicates the target annual number of vehicles served by a single charging station in the region. Indicates the electrification rate of vehicles served by a single charging station in the target year, represents the charging service demand density value, It represents the average number of vehicles served per gas station per year. represents the annual average electrification rate of vehicles served per gas station. Indicates the service range of a single charging station;

[0143] According to the grid division principle, sample grids, the position of the i-th sample grid is , the charging demand is ;pass The most reasonable number k of charging blocks that needs to be further optimized is obtained, and the expression is as follows:

[0144] ;

[0145] in, is the total charging demand in the planning area, and the optimal charging demand for each charging block for:

[0146] ;

[0147] Randomly select k sample grids as the preset initial center points of each charging block area, and the geometric center point of the jth initial center point is used as its coordinate position ;

[0148] S302, Calculate distance and assign grid: distance metric It is a function of spatial distance, which represents the distance between the i-th grid and the j-th center point. The expression is as follows:

[0149] ;

[0150] According to the distance calculation result, the sample grid is assigned to the central grid block with the smallest distance. The expression is as follows:

[0151] ;

[0152] in, represents the charging block index to which the sample grid i is assigned, and arg represents the action corresponding to the optimal strategy;

[0153] S303, update the charging demand of each charging block area: As the sample grid is allocated, the charging demand of each charging block area j is updated. The expression is as follows:

[0154] ;

[0155] in, is the set of sample grids in the jth charging block, For Grid Charging demand;

[0156] S304, update the center position of the charging block area: the center position of the current charging block area is determined by the weighted average position of all small grids in the charging block area, and the update formula is:

[0157] ;

[0158] S305, demand balance: calculate demand error: demand error for each block j Defined as: ;

[0159] Introduce a balancing strategy to ensure that the charging demand of each charging block is close to the target value ;

[0160] Balancing strategy: Adjust the membership of charging blocks based on demand differences, or shift charging service grids from charging blocks with excessive demand to charging blocks with insufficient demand, until the demand deviation reaches an acceptable range.

[0161] Convergence condition: When the demand error of all charging blocks Less than the predetermined threshold When , the algorithm stops iterating, and the expression of the convergence condition is as follows:

[0162] ;

[0163] in, Indicates the maximum allowable value of demand quantity difference;

[0164] S306, output result: When the total demand of each charging block area is close to the target demand amount P, clustering is completed, and k charging blocks with balanced charging demand amounts are output.

[0165] Specifically, in S400, the objective function is to maximize the annual operating income of the charging station in the target charging block, including the charging service fee, V2G discharge response service income, annualized construction cost of the charging station, and annual operating cost. The objective function is expressed as follows:

[0166] ;

[0167] in, is the annualized total return, For charging station operation service income, To convert the construction cost to annual cost, The annual operation and maintenance cost of the charging station;

[0168] The charging station operation service income , the calculation process is as follows:

[0169] ;

[0170] ;

[0171] ;

[0172] in, Indicates the annual charging service fee of the charging station; represents the service revenue of V2G discharge response in one year; The unit price of the service per kilowatt of charging power; represents the service probability coefficient of the station; represents the number of V2G discharges in the region in a year; Indicates the amount of electricity discharged in a single time; represents the service fee for V2G discharge;

[0173] The construction cost of the charging station , the calculation process is as follows:

[0174] ;

[0175] ;

[0176] ;

[0177] ;

[0178] in, represents the average discount rate; Indicates the depreciation period; Indicates the charging stations; Indicates the number of alternative charging stations within the charging block area; represents the infrastructure operation cost; represents the purchase cost of charging piles at the i-th charging station; represents the cost of supporting facilities; The unit price of the land occupied by the station; Indicates the ratio coefficient between the charging power and the floor area of the station; It represents the comprehensive cost per watt of charging infrastructure construction; is the equivalent investment coefficient of supporting facilities cost;

[0179] The operation and maintenance costs of the charging station , the calculation process is as follows:

[0180] ;

[0181] ;

[0182] in, is the annual unit labor cost; is a step function, Indicates the minimum number of fast charging piles at the station, rounded down; Charging infrastructure construction and annual maintenance cost per kilowatt.

[0183] Specifically, in S400 , the constraints include power supply and demand balance constraints, charging station power constraints, and charging station distance constraints;

[0184] The expression of power supply and demand balance constraint is as follows:

[0185] ;

[0186] in, It is The number of alternative charging station sites within the block charging area; For the Block charging: The total amount of charging demand in the block area;

[0187] The expression of the charging station power constraint is as follows:

[0188] ;

[0189] in, Indicates that the total power of the charging piles that can be connected to the station should not exceed the maximum power allowed by the power grid;

[0190] The expression of the charging station distance constraint is as follows:

[0191] ;

[0192] Specifically, in S400, the Aurora optimization algorithm is called to obtain the optimal planning scheme for charging facilities. The process is as follows:

[0193] S401, Initialization: Randomly initialize the charging station power allocation value Initialize the auroral group position and velocity, where each position represents a power allocation scheme and the velocity represents the update rate;

[0194] S402, calculate the objective function: For each charging station power allocation scheme, calculate the objective function:

[0195] ;

[0196] in, Indicates the Total revenue of the station in each block area;

[0197] S403, Aurora group position update: Update the power allocation of each charging station:

[0198] ;

[0199] in, is the optimal location of the current charging station i; For the optimal power allocation solution; and are adjustment factors, used to control the convergence speed;

[0200] S403, check the constraints: after each update, check whether the power allocation meets the constraints. If not, a penalty function is used to impose a cost penalty on the power allocation solution that does not meet the constraints. The penalty cost is The expression is as follows:

[0201] ;

[0202] in: is the penalty factor, which controls the intensity of the penalty; when When , the penalty function value is zero, indicating that the constraint condition is satisfied; otherwise, the penalty cost increases, which makes the benefit value of the objective function smaller and reduces the excellence of the power allocation scheme;

[0203] S404, termination of iteration: When the maximum number of iterations is reached or the objective function converges to a predetermined accuracy, the algorithm is stopped to obtain the optimal power allocation solution for each charging station. .

[0204] Furthermore, based on the optimal charging power of each station and the actual conditions of the station, the specific construction plan and equipment selection plan of the charging station are recommended. The calculation method for the total number of fast guns of public charging infrastructure is:

[0205] ;

[0206] in, Indicates the In the block The total number of public charging infrastructure stations; Indicates rounding up; Indicates the power of a standard fast charging gun.

[0207] In addition, this embodiment also includes a computer device, including a memory and a processor;

[0208] The memory is used to store a computer program that can be executed on the processor;

[0209] The processor is used to implement the steps of the charging facility planning method as described above when executing the computer program.

[0210] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the computer device.

[0211] The computer device may be a mobile phone, desktop computer, laptop, PDA, cloud server, or other computing device. The computer device may include, but is not limited to, a processor and memory. For example, the computer device may also include input and output devices, network access devices, buses, etc.

[0212] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor. The processor is the control center of the computer device and connects the various parts of the entire computer device using various interfaces and lines.

[0213] The memory can be used to store the computer programs and / or modules. The processor implements the computer programs by running or executing the computer programs and / or modules stored in the memory and accessing data stored in the memory. The memory may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as a sound playback function or an image playback function); the data storage area may store data generated based on the use of the mobile phone (such as audio data, a phone book, etc.). Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0214] If the module / unit integrated into the computer device is implemented as a software functional unit and sold or used as a standalone product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can also implement all or part of the process steps in the above-mentioned method embodiments by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a removable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunications signal, and a software distribution medium.

[0215] In addition, an embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned charging facility planning method are implemented.

[0216] This embodiment provides a charging facility planning method based on machine learning. This method utilizes a clustering algorithm to optimize the service radius and range of charging stations. This method effectively addresses the incomplete analysis and irrational optimization zoning issues found in existing technologies. It comprehensively considers all factors influencing charging facility layout, thereby achieving efficient layout and resource allocation for the charging network, significantly improving the overall quality of charging services and user satisfaction. Furthermore, this embodiment utilizes the charging balance equation optimization technology based on the Aurora Optimization Algorithm, effectively shortening optimization time, improving optimization accuracy, and enhancing optimization compatibility.

[0217] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0218] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A charging facility planning method based on machine learning, characterized in that: The following steps are involved: S100: Divide the planning area into grids, including multiple charging service grids with the same service range; S200: Carry out vehicle ownership prediction and charging facility demand prediction in sequence based on the charging service grid to obtain charging demand of the charging service grid; S300: A clustering algorithm is used to perform a fusion analysis of the grids in the planning area. The charging resource conditions and service capabilities of the planning area are comprehensively considered. Based on the charging demand and spatial distribution characteristics of the charging service grid, multiple charging blocks with balanced charging demand are clustered. S400: Construct an objective function based on maximizing the revenue of each charging block, establish constraints based on full coverage of charging demand and optimal utilization of charging resources, invoke the Aurora optimization algorithm to obtain the optimal charging facility planning scheme, and complete the charging facility planning using the optimal charging facility planning scheme. In S400, the Aurora optimization algorithm is called to obtain the optimal planning scheme for charging facilities. The process is as follows: S401, Initialization: Randomly initialize the charging station power allocation value P i Initialize the auroral group position and velocity, where each position represents a power allocation scheme and the velocity represents the update rate; S402, calculate the objective function: For each charging station power allocation scheme, calculate the objective function: max(C total ,j)=P ser (P i )-C inv (P i )-C om (P i )-C penalty (P i ); Among them, C total ,j represents the total revenue of the station in the jth block area; S403, Aurora group position update: Update the power allocation of each charging station: P i (t+1)=P i (t)+α·(P best -P i (t))+β·(P global -P i (t)); Among them, P best is the optimal location of the current charging station i; P global is the optimal power allocation scheme; α and β are adjustment factors used to control the convergence speed; S403, check the constraints: after each update, check whether the power allocation meets the constraints. If not, a penalty function is used to impose a cost penalty on the power allocation solution that does not meet the constraints. The penalty cost C penalty (P i ) is as follows: Among them: λ is the penalty factor, which controls the intensity of the penalty; when When , the penalty function value is zero, indicating that the constraint condition is satisfied; otherwise, the penalty cost increases, which makes the benefit value of the objective function smaller and reduces the excellence of the power allocation scheme; S404, termination of iteration: When the maximum number of iterations is reached or the objective function converges to a predetermined accuracy, the algorithm is stopped to obtain the optimal power allocation solution P for each charging station. i * .

2. The charging facility planning method based on machine learning according to claim 1, characterized in that: In S100 , the charging service grid is a square grid of uniform size, with a side length ranging from 0.5 km to 1 km.

3. The charging facility planning method based on machine learning according to claim 1, characterized in that: In S200, the vehicle ownership prediction includes sequentially performing the car ownership prediction and the electric vehicle ownership prediction; S201, Car Ownership Forecast: For private cars, use the growth rate method to predict the per capita disposable income level in the planning area by the target year. Use the per capita disposable income level to determine the benchmark area. Simultaneously, verify the data using the population age structure and the average number of cars per household method. For other types of cars, the growth rate method is used for forecasting, and data is revised based on the actual development of different types of models; S202, Electric Vehicle Ownership Forecast: Based on the vehicle ownership forecast results, the Bass diffusion model is used to carry out electric vehicle ownership forecast. The cumulative curve expression and density expression of the Bass diffusion model are as follows: Where N(t) represents the cumulative number of electric vehicles adopted before time t; m represents the maximum market potential; p represents the innovation coefficient; q represents the imitation coefficient; and n(t) represents the number of electric vehicle adopters at time t.

4. The charging facility planning method based on machine learning according to claim 3, characterized in that: In S200 , the charging facility demand forecast includes the total charging facility demand forecast and the charging facility demand forecast for each grid; S203, Forecasting the Total Demand for Charging Facilities: Different vehicle-to-charging pile ratios are assigned to different types of vehicles. Based on the number of electric vehicles in use and a differentiated approach, the annual charging demand for each type of electric vehicle is estimated. The total charging power of the charging facilities in the planned area is calculated to complete the forecast of the total demand for charging facilities. S204, forecasting demand for charging facilities in each grid: Based on the actual annual power demand, target demand allocation is made for the total number of charging facilities according to the population, road network density, and refined oil consumption of each grid. The forecast formula for charging facility demand in each grid is as follows: The actual charging demand ratio of a certain grid = the population ratio of a certain grid × 0.4 + the refined oil consumption ratio of a certain grid + 0.4 + the road network density ratio of a certain grid × 0.

2.

5. The charging facility planning method based on machine learning according to claim 4, characterized in that: In S300, the process of using the clustering algorithm to perform fusion analysis on the grids of the planning area is as follows: S301, initialization: Referring to the level of refined oil consumption at gas stations in the region, combined with the forecast results of electric vehicle ownership, matching the charging service capacity of the charging station, and calculating the target year single-station service vehicle, electrification rate, and charging service demand density value. The expression is as follows: b2 = Forecast of electrification rate in target year Where b1 represents the number of vehicles served by a single charging station in the target year, b2 represents the electrification rate of vehicles served by a single charging station in the target year, b3 represents the density of charging service demand, a1 represents the average number of vehicles served by a single gas station in a year, a2 represents the average electrification rate of vehicles served by a single gas station in a year, and S represents the service range of a single charging station. According to the grid division principle, n sample grids are obtained, and the position of the i-th sample grid is (x i ,y i ), the charging demand is p i ; The most reasonable number k of charging blocks that need to be further optimized is obtained through b3, and the expression is as follows: Among them, P all is the total charging demand in the planning area, and the optimal charging demand P for each charging block is: Randomly select k sample grids as the preset initial center points of each charging block area, and the geometric center point of the jth initial center point is used as its coordinate position (x j ,y j ); S302, calculate the distance and assign the grid: The distance metric D(i,j) is a function of the spatial distance, which represents the distance between the i-th grid and the j-th center point. The expression is as follows: According to the distance calculation result, the sample grid is assigned to the central grid block with the smallest distance. The expression is as follows: Among them, j i represents the charging block area to which the sample grid i is assigned, and arg represents the action corresponding to the optimal strategy; S303, update the charging demand of each charging block area: With the allocation of sample grids, the charging demand P of each charging block area j is updated. j The expression is as follows: Among them, O j is the set of sample grids in the jth charging block, f i is the charging demand of grid i; S304, update the center position of the charging block area: the center position of the current charging block area is determined by the weighted average position of all small grids in the charging block area, and the update formula is: S305, demand balance: calculate demand error: demand error Δ for each block j j Defined as: Δ j =|P j -P|; Introduce a balancing strategy to ensure that the charging demand of each charging block is close to the target value P; Balancing strategy: Adjust the membership of charging blocks based on demand differences, or shift charging service grids from charging blocks with excessive demand to charging blocks with insufficient demand, until the demand deviation reaches an acceptable range. Convergence condition: When the demand error of all charging blocks Δ j When it is less than the predetermined threshold ∈, the algorithm stops iterating. The expression of the convergence condition is as follows: in, Indicates the maximum allowable value of demand quantity difference; S306, output result: When the total demand of each charging block area is close to the target demand amount P, clustering is completed, and k charging blocks with balanced charging demand amounts are output.

6. The charging facility planning method based on machine learning according to claim 5, characterized in that: In S400, the objective function is to maximize the annual operating income of the charging station in the target charging block, including the charging service fee, V2G discharge response service income, annualized construction cost of the charging station, and annual operating cost. The objective function is expressed as follows: maxF=S ser -C inv -C om ; Among them, F is the annualized total return, S ser The revenue from charging station operation services, C inv To convert the construction cost to annual cost, C om The annual operation and maintenance cost of the charging station; The charging station operation service income S ser , the calculation process is as follows: S cha,i =D·δ i ·P i ; S V2G =d i ·m·(P j / 60)·h·t; Among them, S cha,i Indicates the annual charging service fee of the charging station; S V2G,i represents the service income of V2G discharge response in one year; D is the service unit price per kilowatt charging power; δ i represents the service probability coefficient of the station; m represents the number of V2G discharges in the area per year; h represents the amount of electricity discharged per time; τ represents the service fee for V2G discharge; The construction cost of the charging station is C inv , the calculation process is as follows: C buy,i =1000×ε×P i ; C sup,i =w sup,i ·P i ; Where r0 represents the average discount rate; y represents the depreciation period; i represents the i-th charging station; I represents the number of alternative charging stations in the charging block area; C basic,i represents the infrastructure operation cost; C buy,i represents the purchase cost of charging piles at the i-th charging station; C sup,i represents the cost of supporting facilities; w i The unit price of the land occupied by the station; represents the ratio coefficient between the charging power and the floor area of the station; ε represents the comprehensive cost per watt of charging infrastructure construction; w sup,i is the equivalent investment coefficient of supporting facilities cost; The charging station operation and maintenance cost C om , the calculation process is as follows: Among them, C labor is the annual unit labor cost; sgn(P i ) is a step function, Indicates the minimum number of fast charging piles at the station, rounded down; C maintain Charging infrastructure construction and annual maintenance cost per kilowatt.

7. The charging facility planning method based on machine learning according to claim 6, characterized in that: In S400 , the constraints include power supply and demand balance constraints, charging station power constraints, and charging station distance constraints; The expression of power supply and demand balance constraint is as follows: Among them, I j is the number of alternative charging stations in the jth charging block area; P j is the total charging demand of the jth charging block; The expression of the charging station power constraint is as follows: 0≤P i ≤P i,Gmax ; Among them, P i,Gmax Indicates that the total power of the charging piles that can be connected to the station should not exceed the maximum power allowed by the power grid; The expression of the charging station distance constraint is as follows: Among them, when P i =0, i =0, indicating that the site is not selected, that is, the service range of the site does not need to be considered; when P i ≠0, i =1, indicating that the site is selected, and the distance between stations must not be less than the minimum value d min .

8. A computer device, characterized in that: including memory and processor; The memory is used to store a computer program that can be executed on the processor; The processor is configured to implement the steps of the charging facility planning method based on machine learning as described in any one of claims 1 to 7 when executing the computer program.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the charging facility planning method based on machine learning as described in any one of claims 1 to 7.

Images