Electric vehicle charging station planning method based on multi-objective optimization

Abstract

The invention discloses an electric vehicle charging station planning method based on multi-objective optimization, and relates to the technical field of electric vehicle charging station planning, and the method comprises the steps: obtaining a target site selection region and a preset planning matching degree, and determining M site selection division sub-regions and M access point sets; determining M correction planning matching degrees; obtaining a power grid security data set of the M access points and a station building economical data set of the M access points; and obtaining a pre-constructed total cost planning function, a voltage offset constraint and a power supply capability margin constraint, performing constraint dual-planning target optimization on the M access point power grid security data set and the M access point station construction economic data set, and determining a target access point. The technical problem that charging reliability and power grid stability cannot be guaranteed due to the fact that electric vehicle charging station planning cannot give consideration to multiple factors in the prior art is solved. The technical effect of improving the planning efficiency and quality of the electric vehicle charging station is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric vehicle charging station planning, and in particular to an electric vehicle charging station planning method based on multi-objective optimization. Background Art

[0002] The widespread adoption of electric vehicles has brought a series of challenges, among which the planning of charging stations is a key issue. The rational planning of charging station access points directly affects charging reliability and grid stability.

[0003] Currently, planning for electric vehicle charging station access points faces numerous shortcomings. For one thing, traditional planning methods often fail to comprehensively consider multiple factors, such as charging reliability and charging demand. This lack of rational charging station access point planning leads to long waiting times for electric vehicles to charge and a poor user experience. Furthermore, some planning methods fail to fully consider the grid's carrying capacity, potentially leading to an over-concentration of access points, resulting in excessive localized grid loads and impacting the grid's safe and stable operation.

[0004] Therefore, researching and developing new methods to rationally plan access points for electric vehicle charging stations has become an urgent issue that needs to be addressed. Summary of the Invention

[0005] The present application provides an electric vehicle charging station planning method based on multi-objective optimization, which is used to solve the technical problem in the prior art that electric vehicle charging station planning cannot take multiple factors into consideration, resulting in the inability to ensure charging reliability and grid stability.

[0006] In view of the above problems, the present application provides an electric vehicle charging station planning method based on multi-objective optimization, the method comprising: Obtain the target site selection area and the preset planning ratio, divide the target site selection area, and determine M site selection sub-areas and M access point sets, where M is a positive integer; Traversing the M site selection sub-areas to perform load disturbance noise iterative analysis, determining M planning uncertainty coefficients based on the analysis results, and revising the preset planning ratio to determine M revised planning ratios; In combination with the M modified planning ratios, according to the preset power grid security index and the preset station construction economic index, traverse the M access point sets to collect data, and obtain the M access point power grid security data set and the M access point station construction economic data set; A pre-built total cost planning function, voltage offset constraint, and power supply capacity margin constraint are obtained, and a constrained dual planning objective optimization is performed on the M access point power grid security data set and the M access point site construction economic data set to determine the target access point.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application obtains a target site selection area and a preset planning ratio, divides the target site selection area, and determines M site selection sub-areas and M access point sets, where M is a positive integer. Iterative load disturbance noise analysis is performed on the M site selection sub-areas, and M planning uncertainty coefficients are determined based on the analysis results. The preset planning ratio is then corrected to determine M corrected planning ratios. Combined with the M corrected planning ratios, data is collected from M access point sets according to preset grid security indicators and preset station construction economic indicators, obtaining M access point grid security data sets and M access point station construction economic data sets. A pre-constructed total cost planning function, voltage offset constraints, and power supply capacity margin constraints are obtained, and dual-planning optimization is performed on the M access point grid security data sets and the M access point station construction economic data sets, respectively, to determine the target access points. This achieves the technical effect of improving the reliability of electric vehicle charging planning site selection and improving the fit of charging station planning with actual needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Attachment Figure 1 This is a flow chart of a method for planning electric vehicle charging stations based on multi-objective optimization provided by an embodiment of the present invention.

[0009] Attachment Figure 2 It is a flow chart of determining M corrected planning ratios in a multi-objective optimization-based electric vehicle charging station planning method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0010] Below in conjunction with specific embodiment, further set forth the present invention.Should be understood that these embodiments are only used to illustrate the present invention and are not used in limiting the scope of the present invention.In addition, should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall equally within the scope limited by the appended claims of the application.

[0011] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0012] Examples, such as the attached Figure 1 As shown, the present application provides an electric vehicle charging station planning method based on multi-objective optimization, wherein the method includes: Obtain the target site selection area and the preset planning ratio, divide the target site selection area, and determine M site selection sub-areas and M access point sets, where M is a positive integer; In one possible embodiment, the target site selection area refers to the city or geographic area of ​​interest during the planning phase, representing the spatial scope for deploying electric vehicle charging stations. The preset planning ratio is the ratio of charging piles to parking spaces when planning electric vehicle charging stations. The M sub-areas for site selection refer to the candidate areas after the target area is divided. The access point set represents the transformer nodes or feeder connection points in the corresponding power grid within each sub-area that can be used to connect to the charging station, with M being the corresponding number of sub-areas and access points.

[0013] First, based on Geographic Information System (GIS) data or urban planning layers, a target area is identified as the research scope for EV charging station construction. Then, combining traffic commuting heat maps and predicted resident travel demand data, the target site selection area is spatially partitioned. For example, a Voronoi diagram-based regional partitioning is used to subdivide the target site selection area into M subareas, ensuring that the partitioning granularity meets planning accuracy requirements. Then, within each subarea, power utility data (such as feeder and transformer GIS layers and SCADA system information) is used to match M access point sets. This generation of M site selection subareas and M access point sets paves the way for subsequent load disturbance analysis and dual-objective optimization. For example, within a 40-square-kilometer urban area, the area can be divided into eight subareas, and at least one access point, such as a substation feeder outlet or a low-voltage switchgear, is identified in each area for subsequent optimization.

[0014] Traversing the M site selection sub-areas to perform load disturbance noise iterative analysis, determining M planning uncertainty coefficients based on the analysis results, and revising the preset planning ratio to determine M revised planning ratios; Further, as attached Figure 2 As shown, the M site selection sub-areas are traversed to perform load disturbance noise iterative analysis, M planning uncertainty coefficients are determined according to the analysis results, and the preset planning ratio is corrected to determine M corrected planning ratios. The steps of the embodiment of the present application also include: Traversing the M selected sub-regions to extract load disturbance data and determine the load disturbance data sequences of the M sub-regions; Traversing the M sub-region load disturbance data sequences to perform multi-scale noise feature analysis to determine the M sub-region load disturbance noise feature sets; Iteratively analyzing the M sub-region load disturbance noise feature sets respectively to determine the M sub-region load disturbance noise iterative features; Performing planning deterministic analysis based on the iterative characteristics of the load disturbance noise in the M sub-areas to determine M planning uncertainty coefficients; The preset planning matching degree is corrected according to the M planning uncertainty coefficients to determine the M corrected planning matching degrees.

[0015] Furthermore, the M sub-region load disturbance noise feature sets are iteratively analyzed to determine M sub-region load disturbance noise iterative features. The embodiment of the present application further includes: Randomly performing non-repetitive combinations of the M sub-region load disturbance noise feature sets in pairs to determine M combination sets; Traversing the M combination sets to construct an element similarity correlation matrix between the load disturbance noise characteristics of two sub-regions in the combination, and determining M element similarity correlation matrix sets; The M element similarity association matrix sets are used to perform feature interaction on the M combination sets, and the interaction results are averaged within the set to determine the iterative features of the load disturbance noise of the M sub-regions.

[0016] Furthermore, the preset planning matching degree is corrected according to the M planning uncertainty coefficients to determine the M corrected planning matching degrees. The embodiment of the present application further includes: Obtaining a preset planning uncertainty coefficient threshold corresponding to the preset planning matching degree; The ratio of the preset planning uncertainty coefficient threshold to the M planning uncertainty coefficients is respectively multiplied by the preset planning matching degree to obtain the M modified planning matching degrees.

[0017] In one possible embodiment, an iterative analysis of the noise affecting load variations in each of the M selected sub-regions is performed to determine the fluctuations in the planning of electric vehicle charging stations for each sub-region. The greater the planning uncertainty coefficient, the greater the uncertainty of the corresponding sub-region as a planning area, and the lower the corresponding planning matching degree, thereby ensuring a stable power supply to the charging stations while ensuring regional electricity consumption. The M revised planning matching degrees reflect the actual matching degrees that meet the requirements of the M sub-regions.

[0018] Preferably, the sub-region load disturbance data sequence is sub-region power load change data recorded in a time series format by a distribution network load monitoring system, reflecting the power demand fluctuations in the region over different time periods. Obtaining these M sub-region load disturbance data sequences provides data support for subsequent noise analysis at different scales, that is, analysis of sub-region load fluctuations.

[0019] A set of multiple analysis scales pre-set by a person skilled in the art is obtained, wherein each analysis scale corresponds to an analysis receptive field, that is, the range of the convolutional network's single data analysis. The multi-analysis scale set is used as the analysis receptive field to construct a set of multi-scale noise feature analyzers. Exemplarily, multiple sample sub-region load disturbance data sequences are obtained, and multiple sample sub-region load disturbance noise feature sets are obtained after performing noise feature extraction according to any analysis scale in the multi-analysis scale set. The framework constructed based on the feedforward neural network is supervised and trained using the multiple sample sub-region load disturbance data sequences and the multiple sample sub-region load disturbance noise feature sets until the training converges, thereby obtaining a trained multi-scale noise feature analyzer. Based on the same principle, the multi-scale noise feature analyzer set is constructed according to different analysis scales. In other words, multi-scale noise feature analysis is a process of decomposing and extracting the changing trend of the sub-region load disturbance data sequence at different time resolutions.

[0020] The M sub-region load disturbance noise feature sets reflect the fluctuation patterns and amplitude structures of the power load in different sub-regions at multiple time levels. Furthermore, by iteratively analyzing the interactions and combinations of features in the M sub-region load disturbance noise feature sets, a stable feature vector is formed, and the M sub-region load disturbance noise iterative features are obtained to determine whether the disturbance characteristics are persistent, repetitive, or intensified.

[0021] Furthermore, the M sub-region load disturbance noise iteration characteristics are analyzed using a planning determinism analyzer pre-constructed by those skilled in the art to obtain the M planning uncertainty coefficients. The planning determinism analyzer is a framework built based on a feedforward neural network and is obtained through supervised training based on sample data. The input data is the sub-region load disturbance noise iteration characteristics, and the output data is the planning uncertainty coefficient.

[0022] Among them, the M planning uncertainty coefficients reflect the predictability and stability of the future load state of the sub-region. The revised planning ratio is a value that is dynamically adjusted according to the uncertainty of the sub-region on the basis of the original preset ratio, and is used to modify the site priority layout weight. For example, if the disturbance standard deviation of a certain area remains greater than 20% under multiple cycles and frequent jumps (>±40%), its uncertainty coefficient may be set to 0.88, while the stable area is only 0.35. Preferably, the preset planning uncertainty coefficient threshold is the maximum value of the regional planning uncertainty coefficient corresponding to the preset planning ratio preset by those skilled in the art. Furthermore, the ratio of the preset planning uncertainty coefficient threshold to the M planning uncertainty coefficients is multiplied by the preset planning ratio to obtain the M revised planning ratios.

[0023] For example, for each of the M selected sub-regions identified in the previous phase, corresponding power load disturbance data sequences are extracted. This power load disturbance data can include SCADA system load curves and charging station time-series power data. The time span is typically 7 to 30 days, with a time resolution of 15 minutes.

[0024] In one possible embodiment, two sub-region load disturbance noise features are randomly selected from the M sub-region load disturbance noise feature sets for combination, and each sub-region load disturbance noise feature appears only once, without repetition or sequence, thereby obtaining the M combination sets. The characteristic element similarity between the two sub-region load disturbance noise features of each combination is calculated using cosine similarity. For example, an element similarity correlation matrix is ​​constructed by calculating the similarity index between the corresponding characteristic components. For example, if the characteristic vectors of the two sub-region load disturbance noise features are: [0.35, 0.42, 0.31] in area A and [0.38, 0.45, 0.29] in area B, the calculated similarity set is: [0.9980.9960.994]. Then, the similarity set is filled into the initially empty matrix to obtain the element similarity correlation matrix. Through the element similarity correlation matrix, the synchronization and difference of the unified sub-region load disturbance patterns under different scales of analysis can be quantified and modeled, thereby providing dynamic feature support for subsequent uncertainty assessment.

[0025] Then, a feature interaction analyzer is obtained, wherein the feature interaction analyzer is used to perform convolution analysis on the element similarity association matrix and the sub-region load disturbance noise characteristics within the combination, thereby determining two enhanced sub-region load disturbance noise characteristics after enhancement, and performing mean calculation on all enhanced sub-region load disturbance noise characteristics within each combination set to obtain M corresponding sub-region load disturbance noise iteration characteristics. Each sub-region load disturbance noise iteration characteristic reflects the load disturbance situation within different sub-regions.

[0026] Preferably, multiple sample element similarity association matrices and multiple sample sub-region load disturbance noise features, as well as corresponding enhanced sub-region load disturbance noise features are obtained as analyzer training data, and the analyzer training data is used to perform supervised training on a framework constructed based on a feedforward neural network until the training converges, thereby obtaining the trained feature interaction analyzer.

[0027] In combination with the M modified planning ratios, according to the preset power grid security index and the preset station construction economic index, traverse the M access point sets to collect data, and obtain the M access point power grid security data set and the M access point station construction economic data set; Furthermore, the preset grid security indicators include annual equipment investment costs, annual distribution network system losses, voltage data and power supply capacity margin, and the preset site construction economic indicators include annual construction costs, annual operation and maintenance costs, land costs and user transfer costs.

[0028] In one possible embodiment, the annual equipment investment cost includes: operating years, distance from the charging station to the access point, line investment cost, and transformer investment cost. The annual network loss cost of the distribution network system includes: real-time power loss. Voltage data includes: voltage upper and lower limits. The power supply capacity margin includes: distribution rated capacity, charging station rated capacity, distribution transformer conventional load, and feeder maximum current carrying capacity. The data that need to be collected for the annual construction cost and the annual operation and maintenance cost include: the number and unit price of transformers in the charging station, the number and unit price of charging piles, capital construction costs, discount rate, and total operating years. The data that need to be collected for land costs include: average land price, area occupied by a parking space, and area occupied by other equipment. The data that need to be collected for user transfer costs include: charging road costs and time costs on the charging road.

[0029] Based on the M revised planned ratios, the charging pile ratio for each access point's electric vehicle charging station is determined. Construction of each access point is then performed according to preset grid security and site construction economic indicators, and data related to grid security is collected. This yields a set of grid security data for each of the M access points and a set of site construction economic data for each of the M access points. This provides basic data for subsequent planning and screening.

[0030] A pre-built total cost planning function, voltage offset constraint, and power supply capacity margin constraint are obtained, and a constrained dual planning objective optimization is performed on the M access point power grid security data set and the M access point site construction economic data set to determine the target access point.

[0031] Furthermore, the embodiment of the present application also includes: A total cost planning function is pre-built, wherein the total cost planning function is: ; in, w 1 is the weight of the planning goal of economically stable operation of the power grid in the total cost planning, w 2 represents the weight of the planning objectives of charging station infrastructure operation and maintenance in the total cost planning. F 1 To achieve the planning goal of economically stable operation of the power grid, F 2 For the planning goal of charging station infrastructure operation and maintenance, Total cost planning function when planning an electric vehicle charging station for an access point; Among them, the planning goal of the economically stable operation of the power grid is F 1 The calculation formula is: , Indicates the weight of the annual cost indicator of equipment investment, Indicates the weight of the annual cost indicator of distribution network system loss, represents the weight of the distribution transformer margin index, is the feeder margin index weight, is the annual cost indicator of equipment investment, is the annual cost of network loss in the distribution network system, is the voltage data, Power supply capacity margin; Among them, the planning objectives of charging station infrastructure operation and maintenance are F 2 The calculation formula is: ; is the weight of the annual construction cost indicator, is the weight of the annual operation and maintenance cost indicator, is the weight of land cost indicator, Transfer cost indicator weights for users, is the annual construction cost, is the annual operation and maintenance cost, is the land cost, Transfer costs to users.

[0032] Furthermore, the embodiment of the present application also includes: The formula for calculating the annual cost of equipment investment is: ; in, represents the discount rate, z represents the operating years, D represents the distance between the charging station and the electrical access point, CL represents the line investment cost per unit distance, and CT represents the transformer investment cost; The formula for calculating the annual cost of distribution network system losses is: ; in, Indicates the conversion coefficient of electricity quantity into electricity price. It represents the increment of power loss of distribution network line at time i; The formula for calculating the distribution transformer margin is: ; in, Indicates the rated capacity of the connected transformer. Indicates the rated capacity of the charging station, Indicates the conventional load of distribution transformer; The formula for calculating feeder margin is: ; in, Indicates the maximum current carrying capacity of the feeder connected to it. Indicates the maximum current of the feeder after connecting to the charging station.

[0033] Furthermore, the embodiment of the present application also includes: ; Among them, e is the number of transformers in the charging station, a is the unit price of the transformer, represents the number of charging piles in the charging station, b represents the unit price of the charging pile, s represents the infrastructure cost of the charging station, represents the discount rate, which refers to the interest rate used to convert future assets into present value and is a parameter reflecting the time value of money used in capital equivalence calculations. z represents the operating life (years), i.e., the total number of years that the charging station will operate and provide charging services after it is built. The formula for calculating the annual operation and maintenance cost is: ; in, is the operation and maintenance factor; The formula for calculating the average annual land cost is: ; Among them, μ represents the land price per square meter (10,000 yuan / ); d represents the area occupied by a complete parking space ( ); Indicates the area occupied by other equipment in the charging station ( ); The formula for calculating user switching costs is as follows: ; in, C H Indicates that charging users are charging from the charging demand point j Average annual travel cost to the charging station, C H1 Indicates that charging users are charging from the charging demand point j The cost of the electricity consumed on the way to the charging station, C H2 Indicates that charging users are charging from the charging demand point j The time cost of traveling to the charging station, J Indicates the number of charging demand points within the service range of the charging station, λ j Indicates charging demand point j The congestion coefficient of the road section to the charging station,d j Indicates the distance from the demand point to the charging station; P 0 means charging electricity price, e Indicates the amount of electricity consumed by electric vehicles per kilometer (kWh / km), f Indicates the conversion factor from time to money, v Indicates that electric vehicles are from the demand point j Average driving speed to the charging station, n j Indicates charging demand point j The average number of electric vehicles that require charging per day.

[0034] In one possible embodiment, the total cost planning function is a weighted sum objective function that combines the grid operation objectives and the construction and operation and maintenance objectives. The voltage deviation constraint refers to the voltage range requirement that must be met after the access point is connected, which is usually a voltage fluctuation within ±5% (such as 0.95~1.05 pu), which is used to prevent voltage anomalies from affecting the operation of the distribution network. The power supply capacity margin constraint refers to the minimum remaining power supply capacity threshold (for example, 10%~30%) that the feeder or distribution transformer must maintain to ensure that subsequent load growth does not cause electrical components to be overloaded. Dual planning objective optimization refers to considering two objective functions ( and ) and the above constraints while performing the process of screening the optimal access point.

[0035] For example, consider a parking lot in a residential area with 120 spaces. The slow charging facilities within the parking lot are located at a fixed location, and the charging service company can decide on their scale. Analyze the selection of electrical access points for different charging infrastructure configurations within the parking lot, considering different objectives.

[0036] The charging pile uses Yangzi Electric's 7kW column-type AC charging pile, with a rated input voltage of AC220V and a rated power of 7kW. The column size is 1430*300*200mm, and the market price is about 3,000 yuan / pile. The simultaneous coefficient K=0.8 of all charging piles during the peak charging period, and the power factor cosψ is 0.95. The total capacity requirement of all charging piles is calculated according to the formula S=KNP / cosψ. Assume that the load capacity requirement of other facilities is 10kVA, where K is the simultaneous coefficient, N is the number of charging piles, and P is the rated power of a single charging pile.

[0037] There are 23 power distribution rooms that can be used as electrical access points. The 0.4kV line uses ZC-VV22-1*70 Cross-linked polyethylene insulated cable, referring to the price quote on the cost-comprehensive website, the unit distance investment cost is about 450 yuan / m, the line operation life is set to 50 years, and the conversion coefficient of electricity consumption into electricity price is set to =0.25; Assuming that the charging service company and the property company have a cooperative relationship and only need to bear the cost of the land area occupied by the charging pile construction, based on the size of the charging pile column, the installation area of ​​a single charging pile is 0.1 The construction cost of the charging station is set at 100,000 yuan.

[0038] Taking into account the changes in the ratio of charging facilities, assuming that in 2025 the charging facilities in the parking lots of this community reach a ratio of 80%, which can fully meet the charging needs of electric private cars in this area, the calculation results of each target under different ratios are shown in Table 1.

[0039] Table 1 Results of various targets under different charging facility ratios in parking lots Ratio Capacity requirement (kVA) <![CDATA[Grid operation and maintenance cost F 1 (yuan)]]> <![CDATA[Construction and operation & maintenance costs of charging stations F 2 (yuan)]]> <![CDATA[Total social cost F 3 (yuan)]]> Access location 40% 308 333 -108842 -650 5 50% 385 388 -141050 -884 5 60% 462 440 -168260 -1078 5 70% 539 904 -195471 -863 11 80% 616 959 -222681 -1053 11 90% 693 1076 -211409 -837 11 From the analysis of Table 1, considering the interests of the power grid company, and comparing the results of F1, it can be seen that when the ratio is 40%, The result is the smallest, then the charging facilities should be configured with a 40% ratio and connected to access point 5. As the ratio increases, the distribution network capacity required by the charging facilities becomes larger and larger, resulting in insufficient capacity margin at access point 5 and the need for expansion. Therefore, access point 11, which is farther away from the parking lot and has a larger margin, is selected to reduce the cost of distribution network transformation. It can be seen that the power grid company prefers to build a smaller scale of charging facilities here, that is, a lower ratio, to ensure the safe operation of the distribution network, minimize the cost of grid transformation for the power grid company, and retain more power supply margin. Considering the interests of the power grid company, compared As a result, it can be seen that when the ratio is 80%, The result is the smallest, then the charging facilities should be configured at an 80% ratio to maximize the benefits for the charging service company. At this time, due to the large demand for charging load capacity, the No. 5 access point, which is closer to the parking lot, has insufficient capacity and needs to be expanded, which requires a large investment. However, the No. 11 access point, which is slightly farther away, has sufficient capacity and does not need to be expanded. Therefore, the No. 11 access point is selected. As a result, it can be seen that when the ratio is 60%, The result is the smallest, so the matching ratio should be 60%. At this time, electrical access point 5 is selected.

[0040] Furthermore, a pre-built total cost planning function, a voltage offset constraint, and a power supply capacity margin constraint are obtained, and a constrained dual planning objective optimization is performed on the M access point power grid security data set and the M access point site construction economic data set to determine the target access point. This embodiment of the application also includes: Mapping and associating the M access point power grid security data sets and the M access point site construction economic data sets respectively to construct M access point associated data particle sets; extracting M first access point associated data particles from the M access point associated data particle sets respectively; Analyzing the fitness values ​​of the M first access point associated data particles under the dual constraints of the voltage offset constraint and the power supply capacity margin constraint using the total cost planning function to determine the fitness values ​​of the M first access point associated data particles; extracting M second access point associated data particles from the set of M access point associated data particles again, and analyzing and determining fitness values ​​of the M second access point associated data particles; determining whether the fitness values ​​of the M first access point-associated data particles are greater than or equal to the fitness values ​​of the second access point-associated data particles; if so, taking directions from the M second access point-associated data particles to the M first access point-associated data particles as M optimization directions, and iterating the M first access point-associated data particles in the set of M access point-associated data particles based on the M optimization directions and a preset optimization step size until a preset number of iterations is satisfied, thereby obtaining M optimal access point-associated data particles; The access points corresponding to the M best access point associated data particles are used as the M best access points, and the access point associated data particle fitness values ​​corresponding to the M best access point associated data particles are used as the M best access point fitness values; The best access point corresponding to the maximum adaptation value of the M best access points is used as the target access point.

[0041] Furthermore, when the fitness values ​​of the M first access point-associated data particles are less than the fitness values ​​of the second access point-associated data particles, directions from the M first access point-associated data particles to the M second access point-associated data particles are used as M optimization directions, and based on the M optimization directions and a preset optimization step size, the M second access point-associated data particles are iterated in the set of M access point-associated data particles until a preset number of iterations is met, thereby obtaining M optimal access point-associated data particles.

[0042] In one possible embodiment, an access point-associated data particle is a particle structure formed by associating and combining grid security data and site economic data for each access point, serving as a candidate solution representation unit in the optimization algorithm. M first access point-associated data particles are then extracted from the set of M access point-associated data particles and used as the starting point for optimization. The total cost planning function is then used, under the dual constraints of voltage offset and power supply capacity margin, to analyze the costs corresponding to the M first access point-associated data particles, i.e., the fitness values ​​of the M first access point-associated data particles, using the aforementioned formulas.

[0043] M second access point associated data particles are extracted again from the set of M access point associated data particles, and the fitness values ​​of the M second access point associated data particles are determined based on the same principle as that of obtaining the fitness values ​​of the M first access point associated data particles.

[0044] Furthermore, when the fitness values ​​of the M first access point-associated data particles are greater than or equal to the fitness values ​​of the second access point-associated data particles, it indicates that the first access point-associated data particles are superior to the second access point-associated data particles. In this case, the directions from the M second access point-associated data particles to the M first access point-associated data particles are used as the M optimization directions. Furthermore, based on a preset optimization step length (i.e., the distance of a single movement) predefined by those skilled in the art, the M first access point-associated data particles are iterated within the set of M access point-associated data particles until a preset number of iterations predefined by those skilled in the art is met. The access point-associated data particles from the last iteration are used as the M best access point-associated data particles. The best access point corresponding to the maximum fitness value among the M best access points is used as the target access point.

[0045] In a possible embodiment, when the fitness values ​​of the M first access point associated data particles are smaller than the fitness values ​​of the second access point associated data particles, it indicates that the second access point associated data particles are superior to the first access point associated data particles. Then, based on the same principle, directions from the M first access point associated data particles to the M second access point associated data particles are used as M optimization directions. The M second access point associated data particles are iterated according to a preset optimization step size until a preset number of iterations is met, and the particle obtained in the last iteration is used as the M best access point associated data particles.

[0046] In summary, the embodiments of the present application have at least the following technical effects: 1. This application forms a total cost planning function by constructing a power grid security index function that integrates the annual equipment investment cost, distribution network system loss, voltage offset and power supply margin, as well as a site construction economic index function that integrates construction cost, operating cost, land cost and user transfer cost. It also introduces dual constraints to fully reflect the optimal site conditions under both technical and economic factors. At the same time, it adopts uncertainty factor correction ratio and particle swarm iterative optimization strategy to achieve a sensitive response of planning results to power grid disturbances and site selection fluctuations, thereby avoiding the problems of unstable access points and low operating efficiency caused by ignoring load fluctuations or power grid constraints in traditional methods, and significantly improving the robustness and reliability of charging station layout in dynamic operating environments.

[0047] 2. This application uses multi-scale noise feature analysis and a random combination iteration mechanism to quantify the load disturbance intensity of each sub-region at different time granularities. It then constructs a similarity correlation matrix and noise iteration features to extract dynamic load interaction characteristics between regions. This uncertainty coefficient is then generated, combined with a preset ratio correction, to achieve dynamic adjustments to site construction trends. This results in improved planning accuracy and zoning management capabilities.

[0048] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0049] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0050] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A multi-objective optimization-based electric vehicle charging station planning method, characterized in that: include: Obtain the target site selection area and the preset planning ratio, divide the target site selection area, and determine M site selection sub-areas and M access point sets, where M is a positive integer; Traversing the M site selection sub-areas to perform load disturbance noise iterative analysis, determining M planning uncertainty coefficients based on the analysis results, and revising the preset planning ratio to determine M revised planning ratios; In combination with the M modified planning ratios, according to the preset power grid security index and the preset station construction economic index, traverse the M access point sets to collect data, and obtain the M access point power grid security data set and the M access point station construction economic data set; A pre-built total cost planning function, voltage offset constraint, and power supply capacity margin constraint are obtained, and a constrained dual planning objective optimization is performed on the M access point power grid security data set and the M access point site construction economic data set to determine the target access point.

2. The electric vehicle charging station planning method based on multi-objective optimization according to claim 1, characterized in that: The preset grid security indicators include the annual equipment investment cost, the annual distribution network system loss cost, voltage data and power supply capacity margin; the preset site construction economic indicators include the annual construction cost, the annual operation and maintenance cost, land cost and user transfer cost.

3. The electric vehicle charging station planning method based on multi-objective optimization according to claim 1, characterized in that: Traversing the M site selection sub-areas to perform load disturbance noise iterative analysis, determining M planning uncertainty coefficients based on the analysis results, and revising the preset planning ratio to determine M revised planning ratios, including: Traversing the M selected sub-regions to extract load disturbance data and determine the load disturbance data sequences of the M sub-regions; Traversing the M sub-region load disturbance data sequences to perform multi-scale noise feature analysis to determine the M sub-region load disturbance noise feature sets; Iteratively analyzing the M sub-region load disturbance noise feature sets respectively to determine the M sub-region load disturbance noise iterative features; Performing planning deterministic analysis based on the iterative characteristics of the load disturbance noise in the M sub-areas to determine M planning uncertainty coefficients; The preset planning matching degree is corrected according to the M planning uncertainty coefficients to determine the M corrected planning matching degrees.

4. The electric vehicle charging station planning method based on multi-objective optimization according to claim 3, characterized in that: Performing iterative analysis on the M sub-region load disturbance noise feature sets respectively to determine the M sub-region load disturbance noise iterative features, including: Randomly performing non-repetitive combinations of the M sub-region load disturbance noise feature sets in pairs to determine M combination sets; Traversing the M combination sets to construct an element similarity correlation matrix between the load disturbance noise characteristics of two sub-regions in the combination, and determining M element similarity correlation matrix sets; The M element similarity association matrix sets are used to perform feature interaction on the M combination sets, and the interaction results are averaged within the set to determine the iterative features of the load disturbance noise of the M sub-regions.

5. The electric vehicle charging station planning method based on multi-objective optimization according to claim 4, characterized in that: Modifying the preset planning matching degree according to the M planning uncertainty coefficients to determine the M modified planning matching degrees includes: Obtaining a preset planning uncertainty coefficient threshold corresponding to the preset planning matching degree; The ratio of the preset planning uncertainty coefficient threshold to the M planning uncertainty coefficients is respectively multiplied by the preset planning matching degree to obtain the M modified planning matching degrees.

6. The electric vehicle charging station planning method based on multi-objective optimization according to claim 1, characterized in that: include: A total cost planning function is pre-built, wherein the total cost planning function is: ； in, w 1 is the weight of the planning goal of economically stable operation of the power grid in the total cost planning, w 2 represents the weight of the planning objectives of charging station infrastructure operation and maintenance in the total cost planning. F 1 To achieve the planning goal of economically stable operation of the power grid, F 2 For the planning goal of charging station infrastructure operation and maintenance, Total cost planning function when planning an electric vehicle charging station for an access point; Among them, the planning goal of the economically stable operation of the power grid is F 1 The calculation formula is: , Indicates the weight of the annual cost indicator of equipment investment, Indicates the weight of the annual cost indicator of distribution network system loss, represents the weight of the distribution transformer margin index, is the feeder margin index weight, is the equipment investment annual cost index, is the annual cost of network loss in the distribution network system, is the voltage data, Power supply capacity margin; Among them, the planning objectives of charging station infrastructure operation and maintenance are F 2 The calculation formula is: ； is the weight of the annual construction cost indicator, is the weight of the annual operation and maintenance cost indicator, is the weight of land cost indicator, Transfer cost indicator weights for users, is the annual construction cost, is the annual operation and maintenance cost, is the land cost, Transfer costs to users.

7. The electric vehicle charging station planning method based on multi-objective optimization according to claim 6, characterized in that: include: The formula for calculating the annual cost of equipment investment is: ； in, represents the discount rate, z represents the operating years, D represents the distance between the charging station and the electrical access point, CL represents the line investment cost per unit distance, and CT represents the transformer investment cost; The formula for calculating the annual cost of distribution network system losses is: ； in, Indicates the conversion coefficient of electricity quantity into electricity price. It represents the increment of power loss of distribution network line at time i; The formula for calculating the distribution transformer margin is: ； in, Indicates the rated capacity of the connected transformer. Indicates the rated capacity of the charging station, Indicates the conventional load of distribution transformer; The formula for calculating feeder margin is: ； in, Indicates the maximum current carrying capacity of the feeder connected to it. Indicates the maximum current of the feeder after connecting to the charging station.

8. The electric vehicle charging station planning method based on multi-objective optimization according to claim 6, characterized in that: include: The formula for calculating the annual construction cost is: ； Among them, e is the number of transformers in the charging station, a is the unit price of the transformer, represents the number of charging piles in the charging station, b represents the unit price of the charging pile, s represents the infrastructure cost of the charging station, represents the discount rate, which refers to the interest rate used to convert future assets into present value and is a parameter reflecting the time value of money used in capital equivalence calculations. z represents the operating life (years), i.e., the total number of years that the charging station will operate and provide charging services after it is built. The formula for calculating the annual operation and maintenance cost is: ； in, is the operation and maintenance factor; The formula for calculating the average annual land cost is: ； Among them, μ represents the land price per square meter (10,000 yuan / ); d represents the area occupied by a complete parking space ( ); Indicates the area occupied by other equipment in the charging station ( ); The formula for calculating user switching costs is as follows: ； in, C H Indicates that charging users are charging from the charging demand point j Average annual travel cost to the charging station, C H1 Indicates that charging users are charging from the charging demand point j The cost of the electricity consumed on the way to the charging station, C H2 Indicates that charging users are charging from the charging demand point j The time cost of traveling to the charging station, J Indicates the number of charging demand points within the service range of the charging station, λ j Indicates charging demand point j The congestion coefficient of the road section to the charging station, d j Indicates the distance from the demand point to the charging station; P 0 means charging electricity price, e Indicates the amount of electricity consumed by electric vehicles per kilometer (kWh / km), f Indicates the conversion factor from time to money, v Indicates that electric vehicles are from the demand point j Average driving speed to the charging station, n j Indicates charging demand point j The average number of electric vehicles that require charging per day.

9. The electric vehicle charging station planning method based on multi-objective optimization according to claim 1, characterized in that: Obtaining a pre-built total cost planning function, a voltage offset constraint, and a power supply capacity margin constraint, performing constrained dual-planning objective optimization on the M access point grid security data set and the M access point site construction economic data set, respectively, to determine the target access point, including: Mapping and associating the M access point power grid security data sets and the M access point site construction economic data sets respectively to construct M access point associated data particle sets; extracting M first access point associated data particles from the M access point associated data particle sets respectively; Analyzing the fitness values ​​of the M first access point associated data particles under the dual constraints of the voltage offset constraint and the power supply capacity margin constraint using the total cost planning function to determine the fitness values ​​of the M first access point associated data particles; extracting M second access point associated data particles from the set of M access point associated data particles again, and analyzing and determining fitness values ​​of the M second access point associated data particles; determining whether the fitness values ​​of the M first access point-associated data particles are greater than or equal to the fitness values ​​of the second access point-associated data particles; if so, taking directions from the M second access point-associated data particles to the M first access point-associated data particles as M optimization directions, and iterating the M first access point-associated data particles in the set of M access point-associated data particles based on the M optimization directions and a preset optimization step size until a preset number of iterations is satisfied, thereby obtaining M optimal access point-associated data particles; The access points corresponding to the M best access point associated data particles are used as the M best access points, and the access point associated data particle fitness values ​​corresponding to the M best access point associated data particles are used as the M best access point fitness values; The best access point corresponding to the maximum adaptation value of the M best access points is used as the target access point.

10. The electric vehicle charging station planning method based on multi-objective optimization according to claim 9, characterized in that: When the fitness values ​​of the M first access point-associated data particles are less than the fitness values ​​of the second access point-associated data particles, directions from the M first access point-associated data particles to the M second access point-associated data particles are used as M optimization directions, and based on the M optimization directions and a preset optimization step size, the M second access point-associated data particles are iterated in the set of M access point-associated data particles until a preset number of iterations is met, thereby obtaining M optimal access point-associated data particles.

Images