Station planning optimization method and system based on charging station data analysis

Abstract

The invention relates to a station planning optimization method and system based on charging station data analysis. The method comprises the following steps: clustering charging record samples; extracting dynamic charging data features of each cluster of user groups, and constructing a dynamic charging data feature time sequence; time sequence modeling is carried out on the dynamic charging data feature sequence through an ARIMA model, regional time period charging demands are predicted according to daily dynamic charging data features of user groups with different charging behaviors, and the predicted regional time period charging demands are weighted and equally divided according to global user density distribution and traffic flow density distribution in the region; obtaining a time period charging demand of each position in the area; obtaining a charging station candidate position set in the region and a target point set in the region; and executing a preset planning strategy to select the candidate position subset of the target charging station to construct the charging station, and meeting the following conditions: all target point sets in the region are covered by at least one charging station, and under the constraint of budget expenditure, the selected charging station position maximally covers the charging demand.

Description

Technical Field

[0001] The present invention relates to the technical field of charging station planning, and in particular, to a method and system for optimizing station planning based on charging station data analysis. Background Art

[0002] Electric vehicles are becoming increasingly popular. However, due to the high construction cost of charging stations and low charging utilization rate, the scale and layout of electric vehicle charging facilities lag far behind the development of electric vehicles, thus restricting the development of electric vehicles. How to reasonably arrange the location, quantity and scale of charging stations is of great significance. Establishing an optimized charging station network is conducive to alleviating the range anxiety of electric vehicle users. Reasonable siting planning of electric vehicle charging stations is also beneficial to improving the service quality and operation efficiency of existing charging facilities. The literature on the research of charging station siting based on population distribution density discloses a charging station siting planning method. Its planning strategy aims at the phenomena of "many vehicles and few charging piles" and "few vehicles and many charging piles" in charging stations. By siting based on population distribution density, more charging stations are built in places with a large population and fewer charging stations are built in places with a small population, so as to avoid the phenomena of "many vehicles and few charging piles" and "few vehicles and many charging piles". The literature focuses on areas such as communities, companies and shopping malls where the population gathers, and determines the population density of each area according to the density of the above-mentioned places in each area, and collects the coordinate points of such areas in the siting area. The coordinates are processed in two steps. First, the coordinate data is subjected to clustering analysis, and similar coordinate points that are clustered in terms of distance are grouped into one category as demand points and charging station alternative points to participate in the charging station siting. The future electric vehicle ownership in the siting area in the near future is predicted through statistical methods, and a model is established and solved with the goal of minimizing the comprehensive cost of charging station construction and electric vehicle searching for stations. The determination of its demand points is more through the clustering of spatial features, without considering the influence of people's charging behaviors, and the planning method is complex. Summary of the Invention

[0003] In order to solve the above technical problems or at least partially solve the above technical problems, the present invention provides a method and system for optimizing station planning based on charging station data analysis.

[0004] In a first aspect, the present invention provides a method for optimizing station planning based on charging station data analysis, including: collecting the geographical location information of charging stations in the area and the daily user order data in the vehicle networking platform to which the charging stations belong, and standardizing and formatting the data of each dimension in all the obtained charging record samples;

[0005] Clustering the charging record samples; extracting the parameters, their means, variances, maximum values and minimum values in the user order data of each time point of each cluster of user groups to obtain the daily dynamic charging data characteristics of different charging behavior user groups, and constructing a time series of dynamic charging data characteristics through historical user order data;

[0006] Perform time series modeling on the dynamic charging data feature sequence through the ARIMA model, predict the charging demand for regional time periods according to the daily dynamic charging data features of different user groups with charging behaviors, and evenly distribute the predicted charging demand for regional time periods weighted by the global user density distribution and traffic flow density distribution within this region, so as to obtain the charging demand for each location within the region during the time period;

[0007] Obtain the set of candidate charging station locations within the region and the set of target points within the region;

[0008] Execute the preset planning strategy to select a subset of candidate charging station locations as the target from the set of candidate charging station locations to construct a charging station, satisfying that all the target points within the region are covered by at least one charging station, and under the constraint of the budget funds, the selected charging station locations maximize the coverage of the charging demand.

[0009] Furthermore, each charging record sample includes: charging power, charging duration, charging frequency, charging station, charging method, charging time period, charging cost;

[0010] Suppose there are n charging record samples, for the i-th sample X i , formatted as X i =(x i1 ,x i2 ,x i3 ,x i4 ,x i5 ,x i6 ,x i7 ), where: x i1 represents the standardized result of the charging power of the i-th sample; x i2 represents the standardized result of the charging duration of the i-th sample; x i3 represents the standardized result of the charging frequency of the i-th sample; x i4 represents the charging station of the i-th sample, where the categorical charging station is transformed into a charging station type code through one-hot encoding; x i5 represents the charging method of the i-th sample, where the categorical charging method is transformed into a charging method type code through one-hot encoding; x i6 represents the standardized result of the charging time period of the i-th sample; x i7 represents the standardized result of the charging cost of the i-th sample.

[0011] Furthermore, the process of clustering the charging record samples includes:

[0012] Initialize K clustering centers:

[0013] C k =(C k1 ,C k2 ,C k3, C k4 , C k5 , C k6 , C k7 ), k = 1, 2, ... K, where C k is the k

[0014] th clustering center, and C kj represents the value on the jth dimension in the kth clustering center;

[0015] The Euclidean distance formula is used to calculate the distance from the charging record sample to the clustering center, and the remaining charging record samples are assigned to the nearest clustering center according to the distance to obtain K clustering clusters;

[0016] Calculate the average value of the charging record samples within the cluster and use it as the clustering center of the cluster. Determine whether the positions of the clustering centers of the K clusters have changed. If they have changed, re-cluster;

[0017] Iterate until the clustering center is stable, complete the clustering, and cluster the charging behaviors in multiple dimensions including charging power, charging duration, charging frequency, charging station, charging method, charging time period, and charging cost.

[0018] Furthermore, randomly select points far apart in the charging record sample dataset as the initial clustering centers; use the elbow method or silhouette coefficient to determine the optimal number of clusters.

[0019] Furthermore, the ARIMA model building process includes:

[0020] Perform the ADF stationarity test on the time series of the dynamic charging data characteristics of each user group; if the data is not stationary, perform differencing processing to make the time series of the dynamic charging data characteristics stationary; use the autocorrelation function and partial autocorrelation function to determine the parameters p, d, q of the ARIMA model; use historical data to train the ARIMA model to model the association between the dynamic charging data characteristics and the regional period charging demand.

[0021] Furthermore, the target points include: the location of the public service area, as well as the midpoints and high-demand areas in the charging station coverage blind spots except for the location of the public service area;

[0022] Among them, identify the charging station coverage blind spots and high-demand areas according to the distribution of charging stations and the distribution of global users in the region:

[0023] Charging station coverage blind spot identification analysis formula:

[0024]

[0025] Among them, U(x, y) is the global user density distribution function, U This the user density distribution threshold, and the blind area covered by the charging station is the position where the global user density distribution function is greater than or equal to the set user density distribution threshold and does not belong to any current charging station;

[0026]

[0027] cs_i is the index of the charging station, and (lat cs_i , lon cs_i ) is the latitude and longitude of the charging station cs_i; r is the coverage radius of the charging station;

[0028] High-demand area identification and analysis formula:

[0029] ,

[0030] The high-demand area represents the area where the user density is higher than the maximum user density distribution within the coverage of all charging stations;

[0031] Among them, U cs_i (x', y') is the density function of the user distribution near the charging station cs_i, and (x', y') is the position coordinate relative to the charging station cs_i.

[0032] Furthermore, the preset planning strategy includes:

[0033] Ignoring the full-coverage constraint of the target point set, under the constraint of the budget, control the selected charging station locations to maximize the coverage of the charging demand, and determine the initial subset of candidate charging station locations;

[0034] Then determine the uncovered target points, and for the uncovered target points, determine the minimum cost required for full coverage;

[0035] Since the minimum cost required for the current full coverage is greater than zero, taking zero here is because the total budget has been spent and the current saved budget obtained by deleting the charging station is zero, and iterate to execute:

[0036] Determine the cost, the proportion of covered target points, and the proportion of provided charging demand of each charging station among all the charging stations corresponding to the current subset of candidate charging station locations. For any charging station, the ratio of the sum of the proportion of covered target points and the proportion of provided charging demand to the cost is used as an evaluation, delete the set number of charging stations with the lowest evaluation, update the current subset of candidate charging station locations, and determine the current saved budget;

[0037] After updating the current subset of candidate charging station locations, determine the uncovered target points, and for the uncovered target points, determine the minimum cost required for the current full coverage;

[0038] Until the minimum cost required for the current full coverage is less than the current saved budget.

[0039] Further, ignoring the full coverage constraint of the target point set D, under the constraint of the budget Cost, control the selected charging station locations to maximize the coverage of the charging demand, and determine the initial subset of candidate charging station locations.

[0040]

[0041] where P h = α∫ A(cs_h) U(x, y)dxdy + β∫ A(cs_h) T(x, y)dxdy, which represents the charging demand covered by the charging station. Since the predicted regional-period charging demand is evenly distributed according to the user density distribution and traffic flow density distribution in this region, the integral results of the user density distribution and traffic flow density distribution weighted by the weights α and β within the coverage area of the charging station can represent the charging demand covered by the charging station; the charging station built at the candidate charging station location h is cs_h, and the coverage area of the charging station cs_h is:

[0042] T(x, y) is the traffic flow density distribution function of the entire region.

[0043] δ h is a selection parameter, taking a value of 0 or 1. When the value is 1, it means building a charging station at the candidate charging station location h; Cost is the total budget, and c h is the cost of building a charging station at the candidate charging station location h.

[0044] In a second aspect, the present invention provides a device for optimizing the station planning based on the analysis of charging station data, including: at least one processing unit, the processing unit is connected to a storage unit through a bus unit, the storage unit stores a computer program, and when the computer program is executed by the processing unit, the method for optimizing the station planning based on the analysis of charging station data as described above is implemented.

[0045] In a third aspect, the present invention provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and is characterized in that when the computer program is executed by a processor, the method for optimizing the station planning based on the analysis of charging station data as described above is implemented.

[0046] The above technical solutions provided by the embodiments of the present invention have the following advantages compared with the prior art:

[0047] The present invention relates to a method and system for optimizing the planning of a charging station based on the analysis of charging station data. The present invention clusters the charging record samples; extracts the dynamic charging data characteristics of each cluster of user groups, and constructs a time series of dynamic charging data characteristics; performs time series modeling on the dynamic charging data characteristic sequence through an ARIMA model, and predicts the charging demand in the regional time period according to the dynamic charging data characteristics of different user groups with charging behaviors. The predicted charging demand in the regional time period is weighted and evenly distributed according to the global user density distribution and traffic flow density distribution in this region to obtain the charging demand at each position in the region; obtains the set of candidate charging station positions in the region and the set of target points in the region; executes a preset planning strategy to select a subset of candidate charging station positions of the target to construct a charging station, satisfying that all the target points in the region are covered by at least one charging station, and under the constraint of the budget, the selected charging station positions maximize the coverage of the charging demand.

[0048] The preset planning strategy includes: ignoring the full coverage constraint of the target point set, under the constraint of the budget, controlling the selected charging station positions to maximize the coverage of the charging demand, and determining an initial subset of candidate charging station positions; then determining the uncovered target points, and for the uncovered target points, determining the minimum cost required for full coverage; since the current minimum cost required for full coverage is greater than zero, taking zero here is because the total budget has been spent and the current saved budget obtained by deleting the charging station is zero, and iteratively execute: determine the cost, the proportion of covered target points, and the proportion of provided charging demand of each charging station among all the charging stations corresponding to the current subset of candidate charging station positions. For any charging station, the ratio of the sum of the proportion of covered target points and the proportion of provided charging demand to the cost is used as an evaluation, delete a set number of charging stations with the lowest evaluation, update the current subset of candidate charging station positions, and determine the current saved budget; after updating the current subset of candidate charging station positions, determine the uncovered target points, and for the uncovered target points, determine the current minimum cost required for full coverage; until the current minimum cost required for full coverage is less than the current saved budget.

[0049] The preset planning strategy gradually iteratively meets the requirements by iteratively deleting low-cost performance candidate positions and introducing a high-cost performance post-rotation value scheme. Through the method of evaluation and screening, the charging stations with low ratios of the proportion of covered target points and the proportion of meeting the charging demand to the cost are deleted, removing the charging stations with low cost performance, saving the budget, and finding a more cost-effective solution. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with the present invention, and are used together with the specification to explain the principles of the present invention.

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0052] Figure 1 It is a flowchart of a method for optimizing station planning based on charging station data analysis provided by an embodiment of the present invention;

[0053] Figure 2 It is a flowchart of clustering charging record samples provided by an embodiment of the present invention;

[0054] Figure 3 It is a flowchart of a preset planning strategy provided by an embodiment of the present invention;

[0055] Figure 4 It is a schematic diagram of a device for optimizing station planning based on charging station data analysis provided by an embodiment of the present invention. Detailed implementation manners

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0057] It should be noted that in this article, the term "including", "comprising", or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements not only includes those elements but also includes other elements not explicitly listed, or further includes elements inherent to such a process, method, article, or device. Without further limitations, an element defined by the statement "including one..." does not exclude the existence of additional identical elements in the process, method, article, or device including the said element.

[0058] Embodiment 1

[0059] As Figure 1 shown, the present application provides a method for optimizing station planning based on charging station data analysis, including:

[0060] Step 1: Collect the geographical location information of charging stations within the collection area and the daily user order data in the vehicle networking platform to which the charging stations belong. The user order data forms a charging record sample, including charging power, charging duration, charging frequency, charging station, charging method, charging time period, and charging cost.

[0061] Standardize and format the data in each dimension of all the obtained charging record samples; assume there are n charging record samples in total, and each charging record sample has 7 dimensions, namely charging power, charging duration, charging frequency, charging station, charging method, charging time period, and charging cost. For the i-th charging record sample X i , it can be formatted as X i =(x i1 , x i2 , x i3 , x i4 , x i5 , x i6 , x i7 ), where: x i1 represents the standardized result of the charging power of the i-th sample; x i2 represents the standardized result of the charging duration of the i-th sample; x i3 represents the standardized result of the charging frequency of the i-th sample; x i4 represents the charging station of the i-th sample, where the categorical charging station is converted into a charging station type code through one-hot encoding; x i5 represents the charging method of the i-th sample, where the categorical charging method is converted into a charging method type code through one-hot encoding; x i6 represents the standardized result of the charging time period of the i-th sample. One example is to divide a day into several time periods, represent them with values corresponding to different time periods, and standardize the data representing different time periods. x i7 represents the standardized result of the charging cost of the i-th sample.

[0062] Step 2: Cluster the charging record samples, extract the parameters, their means, variances, maximum values, and minimum values in the user order data at each time point for each cluster of user groups to obtain the daily dynamic charging data characteristics of different charging behavior user groups, and construct a time series of dynamic charging data characteristics through historical user order data.

[0063] As Figure 2 shown, the clustering process includes: Initialize K clustering centers:

[0064] C k =(C k1 , C k2 , C k3 , C k4 , C k5 , Ck6 , C k7 ), k = 1, 2, ... K, where C k is the k

[0065] th cluster center, and C kj represents the value on the jth dimension in the kth cluster center. In the specific implementation process, randomly select points that are far apart in the charging record sample dataset as the initial cluster centers. Use the elbow method or silhouette coefficient to determine the optimal number of clusters.

[0066] Use the Euclidean distance formula to calculate the distance from the charging record sample to the cluster center, and allocate the remaining charging record samples to the nearest cluster center according to the distance to obtain K clusters.

[0067] Among them, the formula for the Euclidean distance is as follows:

[0068]

[0069] Among them, this formula is used to measure the distance between the ith charging record sample X i and the kth cluster center C k . The smaller the distance, the higher the similarity between the charging record sample X i and the cluster center C k in terms of charging amount, charging duration, charging frequency, charging station, charging method, charging time period, and charging cost.

[0070] The formula for allocating the sample to the nearest cluster center is as follows:

[0071]

[0072] Among them, for each charging record sample X i , by comparing its distance d(X i , C k ) with all K cluster centers, find the cluster center C k with the smallest distance, and then allocate the sample X i to the cluster represented by this cluster center. means taking the k value that makes d(X i , C k ) the smallest, that is, determining the cluster to which the sample X i belongs.

[0073] Calculate the average value of the charging record samples within the cluster and use it as the cluster center of the cluster. Judge whether the positions of the cluster centers of the K clusters have changed. If they have changed, re-cluster;

[0074] Let S k be the set of charging record samples assigned to the kth cluster center, |Sk | is the number of charging record samples in the set S k The formula for updating the cluster center is as follows:

[0075]

[0076] Among them, the formula means that for all charging record samples X assigned to the k-th cluster i The values on any dimension are summed, and then divided by the number of charging record samples |S in the k-th cluster k | to obtain the value of the new cluster center on any dimension. By continuously updating the cluster center, the cluster center can better represent the data characteristics within the cluster.

[0077] Iterate until the cluster center is stable to complete the clustering. If the change in the cluster center position is small, it is considered that the clustering result has converged, and the final result is obtained.

[0078] After clustering, the quality of clustering is evaluated by the mean square error. The mean square error formula is:

[0079]

[0080] Among them, the formula calculates the mean of the sum of the squares of the distances from all charging record samples X i to their respective cluster centers. The smaller the MSE value, the closer the charging record samples are within their respective clusters, that is, the better the clustering effect. n is the total number of charging record samples, represents the square of the distance from the charging record sample X i to the nearest cluster center.

[0081] Cluster the charging behaviors in multiple dimensions including charging power, charging duration, charging frequency, charging station, charging method, charging time period, and charging cost; extract the parameters, their means, variances, maximum values, and minimum values in the user order data at each time point for each cluster of user groups to obtain the daily dynamic charging data characteristics of different charging behavior user groups, and construct a time series of dynamic charging data characteristics through historical user order data.

[0082] In the third step, perform time series modeling on the dynamic charging data characteristic sequence through the ARIMA model, predict the regional time period charging demand according to the daily dynamic charging data characteristics of different charging behavior user groups, and evenly distribute the predicted regional time period charging demand according to the user density distribution and traffic flow density distribution within this region, so as to obtain the time period charging demand at each location within the region.

[0083] In the specific implementation process, the ARIMA model modeling process includes:

[0084] Perform the ADF stationarity test on the time series of the dynamic charging data characteristics for each user group; if the data is not stationary, perform differencing to make the time series of the dynamic charging data characteristics stationary; use the autocorrelation function and partial autocorrelation function to determine the parameters p, d, and q of the ARIMA model; use historical data to train the ARIMA model to model the association between the dynamic charging data characteristics and the charging demand in the regional time period.

[0085] Average the predicted charging demand in the regional time period according to the user density distribution and traffic flow density distribution within the region, so as to obtain the charging demand at each location within the region. In the specific implementation process, obtain the user density distribution and traffic flow density distribution near the charging station and the regional global through the GIS information system.

[0086] Fourth, obtain the set H of candidate locations for charging stations within the region and the set D of target points within the region.

[0087] The set D of target points includes: the location where the public service area is located, as well as the midpoints and high-demand areas in the blind spots covered by the charging stations except for the location where the public service area is located.

[0088] Among them, identify the blind spots covered by the power supply stations and high-demand areas according to the distribution of charging stations within the regional global and the regional global user distribution. Blind spot identification analysis formula for charging station coverage:

[0089]

[0090] Among them, U(x, y) is the user density distribution function of the regional global, U Th is the user density distribution threshold, and the blind spot covered by the charging station is the location where the global user density distribution function is greater than or equal to the set user density distribution threshold and does not belong to any current charging station;

[0091]

[0092] cs_i is the index of the charging station, (lat cs_i , lon cs_i ) is the longitude and latitude of the charging station cs_i; r is the coverage radius of the charging station;

[0093] High-demand area identification analysis formula:

[0094]

[0095] , the high-demand area represents the area where the user density is higher than the maximum user density distribution within the coverage of all charging stations;

[0096] Among them, U cs_i(x', y') is the density function of the user distribution near the charging station cs_i, and (x', y') are the position coordinates relative to the charging station cs_i.

[0097] In the fifth step, execute the preset planning strategy to select a subset of candidate charging station locations as the target from the candidate charging station location set to construct a charging station, satisfying that all target point sets in the area are covered by at least one charging station, and under the budget constraint, the selected charging station locations maximize the coverage of charging demand.

[0098] As Figure 3 shown, the preset planning strategy includes:

[0099] Ignoring the full coverage constraint of the target point set D, under the budget constraint of Cost, control the selected charging station locations to maximize the coverage of charging demand, and determine the initial subset of candidate charging station locations.

[0100]

[0101] Among them, P h = α∫ A(cs_h) U(x, y)dxdy + β∫ A(cs_h) T(x, y)dxdy, which represents the charging demand covered by the charging station. Since the predicted regional time period charging demand is evenly distributed according to the user density distribution and traffic flow density distribution in this area, the integral results of the user density distribution and traffic flow density distribution weighted by the weights α and β in the charging station coverage area can represent the charging demand covered by the charging station; the charging station built at the candidate charging station location h is cs_h, and the coverage area of the charging station cs_h is:

[0102] T(x, y) is the traffic flow density distribution function of the regional global.

[0103] δ h is a selection parameter, taking a value of 0 or 1. Taking a value of 1 means building a charging station at the candidate charging station location h. Cost is the total budget, and c h is the cost of building a charging station at the candidate charging station location h.

[0104] Then determine the uncovered target points, and for the uncovered target points, determine the minimum cost required for full coverage;

[0105]

[0106] Among them, is the initial subset of candidate supplementary charging station locations for the uncovered target points, representing that the target point set is included in the initial subset of candidate supplementary charging station locations. and the initial subset of candidate charging station locations the coverage area of the charging stations built on it

[0107] Since the minimum cost required for full coverage in the initial state is greater than zero, that is, Here, zero is taken because the total budget has been spent, and the initial savings budget obtained by deleting charging stations is zero. The iteration is executed as follows:

[0108] Determine the current subset of candidate charging station locations Among all the corresponding charging stations, for each charging station, calculate the cost, the proportion of covered target points, and the proportion of provided charging demand. For any charging station, use the ratio of the sum of the proportion of covered target points and the proportion of provided charging demand to the cost as an evaluation. Delete the set number of charging stations with the lowest evaluation, and update the current subset of candidate charging station locations

[0109] Determine the current savings budget based on the set number of charging stations with the lowest evaluation deleted:

[0110]

[0111] By means of evaluation and screening, delete the charging stations with low ratios of the proportion of covered target points and the proportion of satisfied charging demand to the cost, remove the charging stations with low cost performance, save the budget, and find a more cost-effective solution.

[0112] After updating the current subset of candidate charging station locations Determine the uncovered target points, and for the uncovered target points, determine the supplementary subset of candidate charging station locations And calculate the minimum cost required for current full coverage:

[0113]

[0114] Until the minimum cost required for current full coverage is less than the current savings budget.

[0115] Example 2

[0116] Refer to Figure 4As shown in the figure, an embodiment of the present invention provides a station planning optimization device based on charging station data analysis, including: at least one processing unit, which is connected to a storage unit through a bus unit. The storage unit, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the software programs, computer-executable programs, and modules corresponding to a method for optimizing station planning based on charging station data analysis in an embodiment of the present invention. The processing unit realizes the above-mentioned method for optimizing station planning based on charging station data analysis by running the software programs, computer-executable programs, and modules stored in the storage unit, including:

[0117] Collect the geographical location information of charging stations in the area and the daily user order data in the vehicle networking platform to which the charging stations belong, and standardize and format the data in each dimension in all the obtained charging record samples;

[0118] Cluster the charging record samples; extract the parameters, their means, variances, maximum values, and minimum values in the user order data at each time point for each cluster of user groups to obtain the daily dynamic charging data characteristics of different charging behavior user groups, and construct a time series of dynamic charging data characteristics through historical user order data;

[0119] Perform time series modeling on the dynamic charging data characteristic sequence through the ARIMA model, predict the regional time period charging demand according to the daily dynamic charging data characteristics of different charging behavior user groups, and evenly distribute the predicted regional time period charging demand according to the global user density distribution and traffic flow density distribution in the region, so as to obtain the time period charging demand at each location in the region;

[0120] Obtain the set of candidate locations for charging stations in the region and the set of target points in the region;

[0121] Execute a preset planning strategy to select a target subset of candidate locations for charging stations from the set of candidate locations for charging stations to construct charging stations, satisfying that all the target points in the region are covered by at least one charging station, and under the constraint of the budget, the selected charging station locations maximize the coverage of the charging demand.

[0122] Of course, the computer program stored in the storage unit of the station planning optimization device based on charging station data analysis provided by the embodiment of the present invention is not limited to the method operations described above, and can also execute the relevant operations in a method for optimizing station planning based on charging station data analysis provided by any embodiment of the present invention.

[0123] Embodiment 3

[0124] An embodiment of the present invention provides a computer-readable storage medium storing a computer program, which when executed, implements the method for optimizing station planning based on charging station data analysis, including:

[0125] Collect the geographical location information of charging stations in the area and the daily user order data in the vehicle networking platform to which the charging stations belong, and standardize and format the data in each dimension in all the obtained charging record samples;

[0126] Cluster the charging record samples; extract the parameters, their means, variances, maximum values, and minimum values in the user order data at each time point for each cluster of user groups to obtain the daily dynamic charging data characteristics of different charging behavior user groups, and construct a time series of dynamic charging data characteristics through historical user order data;

[0127] Perform time series modeling on the dynamic charging data characteristic sequence through the ARIMA model, predict the regional time period charging demand according to the daily dynamic charging data characteristics of different charging behavior user groups, and evenly distribute the predicted regional time period charging demand according to the global user density distribution and traffic flow density distribution in this region, so as to obtain the time period charging demand at each location in the region;

[0128] Obtain a set of candidate locations for charging stations in the area and a set of target points in the area;

[0129] Execute a preset planning strategy to select a target subset of candidate locations for charging stations from the set of candidate locations for charging stations to construct charging stations, satisfying that all the target points in the region are covered by at least one charging station, and under the constraint of the budget, the selected charging station locations maximize the coverage of charging demand.

[0130] The computer program stored in the computer-readable storage medium provided by the embodiment of the present invention is not limited to the method operations as described above, and can also execute the related operations in a method for optimizing station planning based on charging station data analysis provided by any embodiment of the present invention.

[0131] In the embodiments provided by the present invention, it should be understood that the disclosed structures and methods can be implemented in other ways. For example, the structural embodiments described above are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces, and the indirect coupling or communication connection of structures or units can be in an electrical, mechanical or other forms.

[0132] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or may be distributed across multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0133] In addition, each functional unit in various embodiments of the present invention may be integrated into a processing unit, may exist separately as individual physical units, or two or more units may be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0134] The above are only specific embodiments of the present invention, enabling those skilled in the art to understand or implement the present invention. Various modifications to these embodiments will be obvious to those skilled in the art. The general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A station planning optimization method based on charging station data analysis, characterized in that: include: Collect the geographical location information of charging stations in the area and the daily user order data in the Internet of Vehicles platform to which the charging stations belong, and standardize and format the data of each dimension in all the charging record samples obtained; Cluster charging record samples; extract parameters and their mean, variance, maximum, and minimum values ​​from the user order data of each cluster of user groups at each time point to obtain the daily dynamic charging data characteristics of user groups with different charging behaviors, and construct a dynamic charging data feature time series through historical user order data; The dynamic charging data feature sequence is modeled through the ARIMA model, and the regional time period charging demand is predicted according to the daily dynamic charging data characteristics of different charging behavior user groups. The predicted regional time period charging demand is weighted and averaged according to the global user density distribution and traffic flow density distribution in the region, so as to obtain the time period charging demand of each location in the region; Obtain a set of candidate locations of charging stations in the area and a set of target points in the area; The preset planning strategy is executed to select a subset of target charging station candidate locations from the charging station candidate location set to construct charging stations, satisfying the following requirements: all target point sets in the area are covered by at least one charging station, and under the budget constraint, the selected charging station locations maximize the coverage of charging needs.

2. The method for optimizing charging station planning based on charging station data analysis according to claim 1 is characterized in that: Each charging record sample includes: charging power, charging time, charging frequency, charging station, charging method, charging time, and charging fee; Assume there are n charging record samples, for the i-th sample X i , formatted as X i =(x i1 ,x i2 ,x i3 ,x i4 ,x i5 ,x i6 ,x i7 ), where: x i1 represents the standardized result of the charging capacity of the i-th sample; x i2 represents the normalized result of the charging time of the i-th sample; x i3 represents the standardized result of the charging frequency of the i-th sample; x i4 represents the charging station of the i-th sample, where the category-type charging station is converted into the charging station type code through one-hot encoding; x i5 represents the charging method of the i-th sample, where the category-type charging method is converted into the charging method type code through one-hot encoding; x i6 represents the normalized result of the charging period of the i-th sample; x i7 Represents the standardized result of the charging cost of the i-th sample.

3. The method for optimizing charging station planning based on charging station data analysis according to claim 1, characterized in that: The process of clustering charging record samples includes: Initialize K cluster centers: C k =(C k1 ,C k2 ,C k3 ,C k4 ,C k5 ,C k6 ,C k7 ),k=1,2,...K, where C k For the kth Cluster centers, C kj Represents the value of the jth dimension in the kth cluster center; The Euclidean distance formula is used to calculate the distance from the charging record sample to the cluster center. The remaining charging record samples are assigned to the nearest cluster center according to the distance to obtain K clusters. Calculate the average value of the charging record samples within the cluster and use it as the cluster center of the cluster. Determine whether the positions of the cluster centers of the K clusters have changed. If so, re-cluster them. Iterate until the cluster center is stable and clustering is completed, and cluster charging behaviors including charging power, charging time, charging frequency, charging station, charging method, charging time period and charging cost are clustered.

4. The method for optimizing charging station planning based on charging station data analysis according to claim 3 is characterized in that: Randomly select points that are far apart in the charging record sample data set as initial cluster centers; use the elbow rule or silhouette coefficient to determine the optimal number of clusters.

5. The method for optimizing charging station planning based on charging station data analysis according to claim 1 is characterized in that: The ARIMA model building process includes: Perform an ADF stationarity test on the time series of dynamic charging data characteristics of each user group; if the data is not stationary, perform differential processing to make the time series of dynamic charging data characteristics stationary; use the autocorrelation function and partial autocorrelation function to determine the parameters p, d, q of the ARIMA model; use historical data to train the ARIMA model to model the relationship between dynamic charging data characteristics and charging demand in regional time periods.

6. The method for optimizing charging station planning based on charging station data analysis according to claim 1, characterized in that: The target points include: the location of the public service area, and the midpoint of the charging station coverage blind spot and the high-demand area except the location of the public service area; Among them, the power supply station coverage blind spots and high-demand areas are identified based on the distribution of charging stations in the region and the distribution of users in the region: Charging station coverage blind area identification analysis formula: Among them, U(x,y) is the global user density distribution function, U Th is the user density distribution threshold. The charging station coverage blind area is the location where the global user density distribution function is greater than or equal to the set user density distribution threshold and does not belong to any current charging station; cs_i is the index of the charging station, (lat cs_i ,lon cs_i ) is the latitude and longitude of the charging station cs_i; r is the coverage radius of the charging station; High demand area identification analysis formula: ， High demand areas refer to areas where the user density is higher than the maximum user density distribution within the coverage area of ​​all charging stations; Among them, U cs_i (x', y') is the density function of the user distribution near the charging station cs_i, and (x', y') is the position coordinate relative to the charging station cs_i.

7. The method for optimizing charging station planning based on charging station data analysis according to claim 1, characterized in that: The preset planning strategy includes: Ignoring the full coverage constraint of the target point set, under the constraint of budget funds, the selected charging station locations are controlled to maximize the coverage of charging demand and determine the initial subset of candidate charging station locations; Then, the uncovered target points are determined, and for the uncovered target points, the minimum cost required for complete coverage is determined; Since the minimum cost required for full coverage is greater than zero, the zero here is because the total budget has been spent. The current budget saved by deleting the charging station is zero. The iterative execution is: Determine the cost, coverage target point ratio, and charging demand ratio of each charging station among all charging stations corresponding to the current charging station candidate location subset. For any charging station, the ratio of the sum of the coverage target point ratio and the charging demand ratio to the cost is used as the evaluation. Delete a set number of charging stations with the lowest evaluation, update the current charging station candidate location subset, and determine the current saving budget. After updating the current subset of candidate charging station locations, the uncovered target points are determined, and for the uncovered target points, the minimum cost required for full coverage is determined; The minimum cost until full coverage is now less than the current savings budget.

8. The method for optimizing charging station planning based on charging station data analysis according to claim 7 is characterized in that: Ignore the full coverage constraint of the target point set D, and under the constraint of budget cost, control the selected charging station locations to maximize the coverage of charging demand and determine the initial subset of candidate charging station locations. Among them, P h =α∫ A(cs_h) U(x,y)dxdy+β∫ A(cs_h) T(x,y)dxdy represents the charging demand covered by the charging station. Since the predicted regional time period charging demand is evenly divided according to the user density distribution and traffic flow density distribution in the region, the integral result of the user density distribution and traffic flow density distribution weighted by weights α and β in the charging station coverage area can represent the charging demand covered by the charging station. The charging station built at the candidate location h of the charging station is cs_h, and the coverage area of ​​the charging station cs_h is: T(x,y) is the traffic flow density distribution function of the whole region; δ h is a selection parameter, which takes a value of 0 or 1. A value of 1 means that a charging station is built at the candidate location h of the charging station. Cost is the total budget, c h is the cost of building a charging station at the candidate charging station location h.

9. A station planning optimization device based on charging station data analysis, characterized in that: include: At least one processing unit, the processing unit is connected to a storage unit via a bus unit, the storage unit stores a computer program, and when the computer program is executed by the processing unit, the station planning optimization method based on charging station data analysis as described in any of claims 1-8 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the station planning optimization method based on charging station data analysis as described in any of claims 1-8 is implemented.

Images