Charging station demand prediction method, terminal equipment and storage medium

Abstract

The invention relates to a charging station demand prediction method, terminal equipment and a storage medium. The method comprises the following steps: collecting charging station data, POI data and dynamic data in a to-be-researched area; taking the charging station as a center, selecting three areas with different sizes, and constructing corresponding sub-graphs based on POIs contained in the areas; selecting one sub-graph from the three sub-graphs according to the current dynamic data pair, and adjusting the edge of each node in combination with the road congestion condition; obtaining embedded features of the sub-graphs and taking the embedded features as spatial features of the corresponding charging stations; constructing a large graph corresponding to the to-be-studied area by combining the spatial features of all the charging stations; taking the spatial characteristics of the charging stations and the charging times of the previous day as the attributes of the nodes corresponding to the charging stations; and inputting the large graph into the STGCN to obtain the predicted charging times of each charging station in the next day. According to the invention, the accuracy and reliability of charging station demand prediction are improved.

Description

Technical Field

[0001] The present invention relates to the field of charging demand prediction, and in particular to a charging station demand prediction method, terminal equipment and storage medium. Background Art

[0002] With the increasing popularity of electric vehicles, the construction and optimization of charging infrastructure have become crucial. Forecasting demand for charging stations is a key component of the planning and operation of EV charging infrastructure. Accurately forecasting charging demand can help grid operators, charging station operators, and urban planning departments optimize resource allocation, ensuring efficient use of charging facilities and reducing energy consumption and costs.

[0003] As the number of electric vehicles increases, the uncertainty and complexity of charging demand are also increasing. Traditional demand forecasting methods are unable to meet the growing demand, especially in the context of rapid urbanization and frequent land use changes, where the impact of the surrounding building environment on charging station demand is becoming increasingly significant.

[0004] Current charging station demand forecasting methods mainly rely on historical data analysis and simple feature engineering, such as time series analysis, regression models, etc. However, these methods show obvious limitations when dealing with complex spatial relationships and nonlinear dependencies.

[0005] The existing methods currently have the following shortcomings:

[0006] (1) Insufficient use of POI information: Existing methods usually treat surrounding POI information as static label features (such as ONE-HOT coding), ignoring the complex relationships and spatial associations between POIs. In addition, the impact of the diversity and uneven distribution of POI types on charging demand has not been effectively modeled.

[0007] (2) Lack of dynamic adjustment mechanism: Existing methods have difficulty in capturing the dynamic impact of POIs on charging demand. For example, the impact of different types of POIs on charging demand may change significantly under different time periods or weather conditions, and existing methods cannot effectively reflect such dynamic changes.

[0008] (3) Limited expressive power of the model: Traditional models have limited ability to capture nonlinear relationships and high-order interactions, making it difficult to meet prediction needs in complex scenarios. Summary of the Invention

[0009] In order to solve the above problems, the present invention proposes a charging station demand prediction method, a terminal device and a storage medium.

[0010] The specific plan is as follows:

[0011] A method for predicting demand for charging stations includes the following steps:

[0012] S1: Collect charging station data, POI data and dynamic data in the area to be studied;

[0013] S2: For each charging station, based on its geographical location, three regions of different sizes are selected with the charging station as the center, and corresponding subgraphs are constructed based on the POIs contained in each region;

[0014] S3: Select a subgraph corresponding to an area of ​​appropriate size from the three subgraphs based on the current dynamic data, and adjust the edges of each node in the selected subgraph based on the current road congestion between the POIs in the selected subgraph;

[0015] S4: Obtain the embedded features of the sub-graph according to the adjusted sub-graph and use them as the spatial features of the corresponding charging station;

[0016] S5: Combine the spatial characteristics of all charging stations to construct a large graph corresponding to the area to be studied. In this large graph, charging stations are used as nodes, and the distance between two nodes is used to construct edges between them. The spatial characteristics of the charging stations and the number of charging times on the previous day are used as attributes of the nodes corresponding to the charging stations.

[0017] S6: Input the large image into STGCN to obtain the predicted number of charging times for each charging station in the next day.

[0018] Furthermore, charging station data includes: charging station ID, geographic location and number of charges per day; POI data includes: POI type and geographic location; dynamic data includes: real-time weather data, holiday data for the day, current time data and real-time traffic flow data.

[0019] Furthermore, the area selected in step S2 is circular, and the corresponding different sizes are different radii, which are 500m, 1km, and 2km respectively.

[0020] Furthermore, when constructing a subgraph, the nodes in the subgraph represent POIs, and the attributes of the nodes adopt the features of the POIs, which are composed of both category features and spatial features.

[0021] Furthermore, a method for selecting a sub-graph corresponding to an area of ​​appropriate size from the three sub-graphs based on the current dynamic data is as follows: initially, the area size of the sub-graph to be selected is set to the middle value of the three sizes, and then the direction is adjusted according to the area sizes corresponding to different types of dynamic data to obtain the final area size of the sub-graph to be selected.

[0022] Furthermore, the region resizing directions corresponding to different types of dynamic data include:

[0023] (1) Determine whether the current weather conditions are suitable for travel based on real-time weather data. If so, reduce the size of the sub-graph area to be selected.

[0024] (2) Determine whether the current time is a peak travel time based on the holiday data of the day and / or the current time data. If so, increase the size of the area to be selected as the subgraph;

[0025] (3) Determine whether the average vehicle speed of the current road is less than the speed threshold based on the real-time traffic flow data. If so, increase the size of the area where the sub-graph needs to be selected.

[0026] Furthermore, in combination with the current road congestion between the POIs in the selected subgraph, the method for adjusting the edges of the nodes in the selected subgraph is as follows: when there is an edge connecting two nodes and the corresponding road changes from unobstructed to congested, the edge connecting the two nodes is canceled, and when the road changes from congested to unobstructed again, the edge between the two nodes is rebuilt.

[0027] Furthermore, the subgraph embedding features are obtained through GraphSAGE.

[0028] A charging station demand prediction terminal device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method described above are implemented in an embodiment of the present invention.

[0029] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method described above in an embodiment of the present invention.

[0030] The present invention adopts the above technical solution to improve the accuracy and reliability of charging station demand prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Shown is a flow chart of embodiment 1 of the present invention. DETAILED DESCRIPTION

[0032] To further illustrate various embodiments, the present invention provides accompanying drawings. These drawings form part of the present disclosure and are primarily used to illustrate the embodiments and, in conjunction with the relevant description in the specification, to explain the operating principles of the embodiments. By referring to these drawings, those skilled in the art will be able to understand other possible implementations and the advantages of the present invention.

[0033] The present invention will now be further described with reference to the accompanying drawings and specific embodiments.

[0034] Example 1:

[0035] The embodiment of the present invention provides a method for predicting demand for charging stations. Figure 1 As shown, the method includes the following steps:

[0036] S1: Collect charging station data, POI data and dynamic data in the area to be studied.

[0037] The charging station data collected in this embodiment include: charging station ID, geographic location and number of charges per day; the POI data collected include: POI type (such as school area, commercial area, residential area, etc.) and geographic location; the dynamic data collected include: real-time weather data, holiday data of the day, current time data and real-time traffic flow data (which can be obtained through the API of the electronic map), etc. In specific implementation, those skilled in the art can adjust the dynamic data to be used as needed.

[0038] Furthermore, this embodiment also includes feature extraction of the collected POI data. In this embodiment, the features of the POI data are set to consist of both category features and spatial features. The category features are obtained by encoding the type of POI, such as using one-hot encoding or label encoding, and converted into a numerical form that can be processed by the model; the spatial features are calculated based on the geographical location of the POI, through the road network, and the navigation distance, direction and other spatial information between the POI and the charging station are calculated, and the spatial features are obtained based on the spatial information.

[0039] S2: For each charging station, based on its geographical location, three regions of different sizes are selected with the charging station as the center, and corresponding subgraphs are constructed based on the POIs contained in each region.

[0040] In this embodiment, the area selected is circular, and the corresponding sizes are different radii, namely 500m, 1km, and 2km, to cover POIs at short, medium, and long distances. In specific implementation, the sizes of the three areas can be specifically selected according to actual applications.

[0041] The nodes in the constructed subgraph are constructed based on the POIs contained in the corresponding area of ​​the subgraph. The attributes of the nodes adopt the characteristics of the POIs, and the edges between the nodes are constructed based on factors such as the distance between the POIs and the road network (that is, an edge between the nodes corresponding to the two POIs is constructed when there is a road connecting the two POIs and the distance between them is less than the distance threshold).

[0042] S3: Select a subgraph corresponding to an area of ​​appropriate size from the three subgraphs according to the current dynamic data, and adjust the edges of each node in the selected subgraph in combination with the current road congestion between the POIs in the selected subgraph.

[0043] When selecting the three subgraphs based on the current dynamic data, different selection methods need to be used according to the type of dynamic data. Specifically, they can be:

[0044] (1) Based on real-time weather data, determine whether the current weather conditions are rainy, snowy, typhoon-like, or other weather conditions that affect travel. If so, reduce the size of the area that needs to be selected so that the user can park the car at the charging station closest to the destination to charge.

[0045] (2) Determine whether the current time is a peak travel time (such as May Day, National Day, morning and evening rush hours, etc.) based on the holiday data of the day and / or the current time data. If so, increase the size of the area to be selected in the sub-graph to expand the range of charging stations for users to choose from;

[0046] (3) Based on the real-time traffic flow data, determine whether the average vehicle speed of the current road is less than the speed threshold (the speed threshold corresponds to the boundary value between congestion and smooth traffic). If so, it means that the current road is congested, and the size of the area that needs to be selected in the sub-graph is increased to facilitate users to expand the range of charging station selection.

[0047] During the specific selection, this embodiment initially sets the area size of the sub-image to be selected to the middle value among the three sizes, and then adjusts the direction (increase or decrease) of the area size corresponding to different types of dynamic data to obtain the final area size of the sub-image to be selected.

[0048] After completing the subgraph selection, it is also necessary to dynamically adjust the edges of each node in the selected subgraph based on the current road congestion between each POI in the subgraph. The specific adjustment method is: when there is an edge connecting two nodes and the corresponding road changes from unobstructed to congested (which can be obtained based on electronic map data, and congestion is usually displayed in red on electronic maps) (because the road between the two nodes may be long, it may only partially change from unobstructed to congested, in which case it is also determined that the corresponding road changes from unobstructed to congested), cancel the edge connecting the two nodes, and reconstruct the edge between the two nodes when the road changes from congested to unobstructed again.

[0049] S4: Obtain the embedded features of the sub-graph according to the adjusted sub-graph and use them as the spatial features of the corresponding charging station.

[0050] In this embodiment, GraphSAGE (Graph Sample and Aggregate) is used to obtain subgraph embedding features. GraphSAGE is a graph neural network algorithm that generates node embedding representations by sampling and aggregating information about neighboring nodes. It is suitable for inductive learning tasks, i.e., it can generate node representation vectors on unseen graphs. Other models can also be used in other embodiments, and are not limited here.

[0051] S5: Combine the spatial characteristics of all charging stations to construct a large graph corresponding to the area to be studied; in the large graph, the charging stations are used as nodes, and the distance between the two nodes corresponding to the charging stations is used to construct edges between the nodes (that is, an edge is constructed when the distance is greater than the preset distance threshold, otherwise no edge is constructed), and the spatial characteristics of the charging station and the number of charging times on the previous day are used as attributes of the node corresponding to the charging station.

[0052] S6: Input the large graph into STGCN (Spatiotemporal Graph Convolutional Network) to obtain the predicted number of charging times for each charging station in the next day.

[0053] The next day here is the second day relative to the previous day, that is, the number of charging times on the second day is predicted based on the number of charging times on the first day.

[0054] This embodiment of the present invention improves upon existing solutions by embedding dynamic multi-scale subgraphs, simplifying POI features into a single feature. This approach considers some of the impact of POIs, but also uniformly treats the impact of the same geographic region, ignoring spatial heterogeneity, real-time traffic flow changes, and multi-scale correlations. For example, of two charging stations within a 1km radius of a primarily residential area, the one with a closer navigation distance and less traffic in the vicinity should be more affected by the residential area, whereas existing solutions ignore this information. This embodiment focuses on the differential impact within the same feature area, constructing multi-granularity subgraphs (500m / 1km / 2km) to achieve in-depth modeling of surrounding POI information. It dynamically adjusts the subgraph range based on real-time traffic flow and weather events (e.g., expanding to 2km in the event of congestion), breaking through the fixed radius limit. It considers the impact of traffic flow and navigation distance, capturing the complex relationships and spatial dependencies between POIs and improving the richness of feature representation. It also uses road congestion (rather than Euclidean distance) to define edge weights, making them more consistent with actual user behavior.

[0055] Example 2:

[0056] The present invention also provides a charging station demand prediction terminal device, including a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, the steps in the above-mentioned method embodiment of embodiment 1 of the present invention are implemented.

[0057] Furthermore, as an executable solution, the charging station demand prediction terminal device can be a computing device such as a desktop computer, laptop, PDA, or cloud server. The charging station demand prediction terminal device may include, but is not limited to, a processor and a memory. Those skilled in the art will understand that the above-mentioned component structure of the charging station demand prediction terminal device is merely an example of a charging station demand prediction terminal device and does not constitute a limitation on the charging station demand prediction terminal device. The charging station demand prediction terminal device may include more or fewer components than those described above, or a combination of certain components, or different components. For example, the charging station demand prediction terminal device may also include input and output devices, network access devices, buses, etc., which are not limited in this embodiment of the present invention.

[0058] Furthermore, as an executable solution, the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices. The general-purpose processor may be a microprocessor or any conventional processor. The processor serves as the control center of the charging station demand forecasting terminal device, connecting various components of the entire charging station demand forecasting terminal device using various interfaces and lines.

[0059] The memory can be used to store the computer programs and / or modules. The processor implements the various functions of the charging station demand prediction terminal device by running or executing the computer programs and / or modules stored in the memory and calling the data stored in the memory. The memory can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required for a function; the data storage area can store data created based on the use of the mobile phone, etc. In addition, the memory can include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0060] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above method in the embodiment of the present invention are implemented.

[0061] If the module / unit integrated in the charging station demand prediction terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, it can implement the steps of the above-mentioned method embodiments. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM) and software distribution medium, etc.

[0062] Although the present invention has been particularly shown and described in conjunction with preferred embodiments, it will be understood by those skilled in the art that various changes in form and details may be made to the present invention without departing from the spirit and scope of the invention as defined in the appended claims, and all such changes are within the scope of protection of the present invention.

Claims

1. A charging station demand forecasting method, characterized in that: The following steps are involved: S1: Collect charging station data, POI data and dynamic data in the area to be studied; S2: For each charging station, based on its geographical location, three regions of different sizes are selected with the charging station as the center, and corresponding subgraphs are constructed based on the POIs contained in each region; S3: Select a subgraph corresponding to an area of ​​appropriate size from the three subgraphs based on the current dynamic data, and adjust the edges of each node in the selected subgraph based on the current road congestion between the POIs in the selected subgraph; S4: Obtain the embedded features of the sub-graph according to the adjusted sub-graph and use them as the spatial features of the corresponding charging station; S5: Combine the spatial characteristics of all charging stations to construct a large graph corresponding to the area to be studied. In this large graph, charging stations are used as nodes, and the distance between two nodes is used to construct edges between them. The spatial characteristics of the charging stations and the number of charging times on the previous day are used as attributes of the nodes corresponding to the charging stations. S6: Input the large image into STGCN to obtain the predicted number of charging times for each charging station in the next day.

2. The charging station demand forecasting method according to claim 1, characterized in that: Charging station data includes: charging station ID, geographic location and number of charges per day; POI data includes: POI type and geographic location; dynamic data includes: real-time weather data, holiday data for the day, current time data and real-time traffic flow data.

3. The charging station demand forecasting method according to claim 1, characterized in that: The area selected in step S2 is circular, and the corresponding different sizes are different radii, which are 500m, 1km, and 2km respectively.

4. The charging station demand forecasting method according to claim 1, characterized in that: When constructing a subgraph, the nodes in the subgraph represent POIs, and the attributes of the nodes adopt the features of the POIs, which are composed of both category features and spatial features.

5. The method for predicting charging station demand according to claim 1, wherein: The method for selecting a sub-graph corresponding to an area of ​​appropriate size from the three sub-graphs according to the current dynamic data is as follows: initially, the area size of the sub-graph to be selected is set to the middle value of the three sizes, and then the direction is adjusted according to the area sizes corresponding to different types of dynamic data to obtain the final area size of the sub-graph to be selected.

6. The charging station demand forecasting method according to claim 5, characterized in that: The area resizing directions corresponding to different types of dynamic data include: (1) Determine whether the current weather conditions are suitable for travel based on real-time weather data. If so, reduce the size of the subgraph area to be selected. (2) Determine whether the current time is a peak travel time based on the holiday data of the day and / or the current time data. If so, increase the size of the area to be selected as the subgraph; (3) Determine whether the average vehicle speed of the current road is less than the speed threshold based on the real-time traffic flow data. If so, increase the size of the area where the sub-graph needs to be selected.

7. The charging station demand forecasting method according to claim 1, characterized in that: Combined with the current road congestion between the POIs in the selected subgraph, the method for adjusting the edges of the nodes in the selected subgraph is as follows: when there is an edge connecting two nodes and the corresponding road changes from unobstructed to congested, the edge connecting the two nodes is canceled, and when the road changes from congested to unobstructed again, the edge between the two nodes is rebuilt.

8. The method for predicting charging station demand according to claim 1, wherein: The embedding features of the subgraph are obtained through GraphSAGE.

9. A charging station demand forecasting terminal device, characterized by: The method comprises a processor, a memory, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium storing a computer program, wherein: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Images