A method and system for analyzing the spatiotemporal characteristics of cluster charging load

Abstract

This invention proposes a method and system for analyzing the spatiotemporal characteristics of clustered charging loads, belonging to the field of electric vehicle technology. The method acquires spatial and charging information of charging pile clusters, then obtains observation data samples based on these information. An improved mean clustering algorithm is then used to cluster the observation data samples. Finally, the spatiotemporal characteristics of the clustered charging load are analyzed based on the location information of each charging pile or subgroup and the clustering results. The improved mean clustering algorithm used in this invention, by setting a class boundary adjustment coefficient, correlates the class boundary adjustment coefficient with the sample data, thus balancing the accuracy and complexity of the clustering process.

Description

Technical Field

[0001] This invention belongs to the field of electric vehicle technology, and in particular relates to a method and system for analyzing the spatiotemporal characteristics of cluster charging load. Background Technology

[0002] The rapid development of the electric vehicle industry has driven the widespread construction of charging piles and other infrastructure. However, the disorderly and irregular charging of electric vehicles in the market has exacerbated the peak-valley load difference in the power grid, especially during peak charging periods at night, which may cause local overload, affecting voltage stability and equipment lifespan. It may also lead to a reduction in the load capacity of some parts of the distribution network, making it unable to meet the needs of other loads. Spatiotemporal characteristic analysis can identify the spatiotemporal distribution patterns of loads, providing a basis for orderly charging strategies and distribution network expansion. It is an important means to solve the load stability and power supply stability problems caused by electric vehicle charging.

[0003] Currently, there are many shortcomings in the analysis of the spatiotemporal characteristics of charging load. For example, Chinese invention patent application CN113077122A provides a method for assessing the spatiotemporal distribution of charging load based on electric vehicles. It first collects electric vehicle charging load data for each charging station, then extracts the morphological features of the charging load curves, obtains the Euclidean distance between the charging load curves, and uses a spectral clustering algorithm to cluster the charging loads based on the morphological features and Euclidean distances to obtain the spatial distribution data of the charging load. Then, it uses a Markov algorithm to calculate the temporal variation data of the charging load, and combines the spatial distribution data and the temporal variation data of the charging load to calculate the charging load characteristics of electric vehicles in different areas and at different times. Chinese invention patent application CN114862033A discloses a method for predicting the spatiotemporal distribution of electric vehicle charging load. It comprehensively considers distance and energy consumption to perform multi-mode clustering of functional areas, calculates the travel OD matrix of each area at each time, and establishes a spatiotemporal distribution model of electric vehicle charging load based on the kernel density estimation method of Gaussian function. This can provide a model basis for subsequent orderly charging strategies for electric vehicles and the selection of charging station locations and capacity formulation. However, neither of the above two methods takes into account the class boundary problem of clustering algorithms, which may lead to low accuracy of clustering results or complex clustering process and low clustering efficiency. Summary of the Invention

[0004] To address the aforementioned problems in the existing technology, this invention proposes a method and system for analyzing the spatiotemporal characteristics of cluster charging loads, thereby solving the problems of low clustering accuracy or low clustering efficiency caused by the division of clustering boundaries in the existing technology.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: a method for analyzing the spatiotemporal characteristics of cluster charging load, comprising the following steps: Step S1: acquiring spatial information and charging information of the charging pile cluster; wherein, the spatial information of the charging pile cluster includes the location information of each charging pile or subgroup of charging piles in the charging pile cluster, and the charging information of the charging pile cluster includes the time-output power curve and charging mode information of each charging pile or subgroup of charging piles in the charging pile cluster; Step S2: obtaining observation data samples based on the time-output power curve and charging mode information of each charging pile or subgroup of charging piles in the charging pile cluster obtained in Step S1; Step S3: clustering the observation data samples in Step S2 using an improved mean clustering algorithm based on class boundary adjustment coefficient, and obtaining clustering results; Step S4: performing spatiotemporal characteristic analysis of cluster charging load based on the location information of each charging pile or subgroup of charging piles in Step S1 and the clustering results in Step S3.

[0006] Furthermore, in step S2, the charging method information includes slow charging and fast charging.

[0007] Further, in step S2, the observed data sample includes multiple sample data Di, where i represents the sample data sequence number, and Di = [X...]. i1 X i2 X i3 ], where X i1 X represents the output time period of the i-th sample data. i2 This indicates that the i-th sample data is in X. i1 Charging method during the output period, X i3 This indicates that the i-th sample data is in X. i1 Output power during the output period.

[0008] Further, in step S3, the improved mean clustering algorithm includes: Step S31: Based on the load type, randomly select M sample data from the observation data samples obtained in step S2 as cluster center data to obtain M classes; where M represents the total number of load types; Step S32: Calculate the distance between each sample data of other sample data in the observation data samples and the M cluster center data, and assign the sample data with the smallest distance to the class corresponding to the corresponding cluster center data; Step S33: For the calculated M classes, calculate the mean of the samples in each class as the new cluster center data, and determine whether the iteration termination condition is met. If it is met, output the calculated M classes as the cluster structure. If it is not met, repeat steps S32-S33 until the iteration termination condition is met.

[0009] Furthermore, in step S32, the formula for calculating the distance between the sample data and the cluster center data is as follows:

[0010]

[0011] Among them, D i,c X represents the distance between sample data i and cluster center data c. i1 X represents the output time period of the i-th sample data. i2 This indicates that the i-th sample data is in X. i1 Charging method during the output period, X i3 This indicates that the i-th sample data is in X. i1 Output power during the output time period, X c1 X represents the output time period of cluster center data c. c2 This indicates that the cluster center data c is in X. c1 The charging method during the output time period, Xc3 represents the cluster center data c in X c1 Output power during the output period.

[0012] Furthermore, in step S33, the cost function corresponding to the clustering result is calculated, and the iteration termination condition is satisfied when the cost function corresponding to the clustering result is minimized.

[0013] Furthermore, the cost function corresponding to the clustering results is calculated according to the following formula:

[0014]

[0015] Where CF represents the cost function corresponding to the clustering result, i represents the sample data index, T represents the total number of sample data, c represents the cluster center data, M represents the total number of load types, and d ic P represents the distance from sample data i to cluster center data c. ic This indicates the probability that sample data i is clustered into the class corresponding to cluster center data c;

[0016] The P ic Obtained from the following formula:

[0017]

[0018] Where ρ represents the class boundary adjustment coefficient, d i1 d represents the distance from sample data i to the first cluster center data. i2 d represents the distance from sample data i to the second cluster center data. iM This represents the distance from sample data i to the Mth cluster center data.

[0019] Furthermore, the class boundary adjustment coefficient ρ is proportional to the total number of sample data T, i.e., ρ = β × T, where β represents the proportionality coefficient.

[0020] This invention also proposes a cluster charging load spatiotemporal characteristic analysis system for performing any of the above-mentioned cluster charging load spatiotemporal characteristic analysis methods. The analysis system includes a data acquisition unit and a data analysis unit, wherein the data acquisition unit is connected to the data analysis unit. The data acquisition unit is used to acquire spatial information of the charging pile cluster and charging information of the charging pile cluster. The spatial information of the charging pile cluster includes the location information of each charging pile or subgroup of charging piles in the charging pile cluster, and the charging information of the charging pile cluster includes the time-output power curve and charging mode information of each charging pile or subgroup of charging piles in the charging pile cluster. The data analysis unit is used to perform spatiotemporal characteristic analysis of the cluster charging load.

[0021] The beneficial technical effect of this invention compared with the prior art is that by setting the class boundary adjustment coefficient, the class boundary adjustment coefficient is associated with the sample data, which can balance the accuracy of clustering and the complexity of the clustering process. Attached Figure Description

[0022] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0023] Figure 1 A flowchart of a method for analyzing the spatiotemporal characteristics of cluster charging load provided by the present invention;

[0024] Figure 2 The flowchart of an improved mean clustering algorithm provided by the present invention is shown. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] The concepts involved in this application will first be described with reference to the accompanying drawings. It should be noted that the following descriptions of various concepts are only for the purpose of making the content of this application easier to understand and do not constitute a limitation on the scope of protection of this application; furthermore, the embodiments and features in the embodiments of this application can be combined with each other unless otherwise specified. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0027] Example

[0028] Combination Figure 1 This invention proposes a method for analyzing the spatiotemporal characteristics of cluster charging load, comprising the following steps:

[0029] Step S1: Obtain the spatial information and charging information of the charging pile cluster; wherein, the spatial information of the charging pile cluster includes the location information of each charging pile or subgroup of charging piles in the charging pile cluster, and the charging information of the charging pile cluster includes the time-output power curve and charging method information of each charging pile or subgroup of charging piles in the charging pile cluster.

[0030] Step S2: Based on the time-output power curves and charging method information of each charging pile or subgroup of charging piles in the charging pile cluster in Step S1, obtain observation data samples;

[0031] Step S3: Cluster the observed data samples in Step S2 using the improved mean clustering algorithm based on class boundary adjustment coefficients to obtain the clustering results;

[0032] Step S4: Based on the location information of each charging pile or charging pile subgroup in Step S1 and the clustering results in Step S3, perform spatiotemporal characteristic analysis of cluster charging load, find charging clusters with high similarity of charging pile loads in different locations, and be able to analyze the composition of different cluster charging loads at the city-district level based on geographical location, network topology and other information, realize the geographical source of charging load at any time, and further improve the analysis level of the distribution of different types of charging loads and their subsequent characteristic evolution.

[0033] The charging method information in step S2 above includes slow charging and fast charging. Slow charging is a low-power, long-time charging method, which usually ensures the battery is fully charged; fast charging is a high-power, short-time charging method, which usually cannot ensure the battery is fully charged. Meanwhile, the observation data samples in step S2 include multiple sample data Di, where i represents the sample data sequence number, Di = [X...]. i1 X i2 X i3 ], where X i1 X represents the output time period of the i-th sample data. i2 This indicates that the i-th sample data is in X. i1 Charging method during the output period, X i3 This indicates that the i-th sample data is in X. i1Output power during the output period. It is understood that both the output period and charging method can be represented digitally. For example, the output period can be set as a two-digit number, with 11, 12, and 13 representing the daily peak period, daily average period, and daily valley period, respectively; the charging method can be represented by a one-digit number, such as 0 and 1 representing slow charging and fast charging, respectively. This digital setting is for the convenience of calculating the mean during subsequent clustering calculations.

[0034] Combination Figure 2 In step S3 above, the specific steps for improving the mean clustering algorithm include:

[0035] Step S31: Based on the load type, randomly select M sample data as cluster center data from the obtained observation data sample to obtain M classes; where M represents the total number of load types. Different battery types can be selected as load types according to the analysis needs. For example, the typical charging power range of ternary lithium batteries is about 300-600W, the typical charging power range of lithium iron phosphate batteries is about 200-300W, and the typical charging power range of lead-acid batteries is about 120-180W. Therefore, the charging load type can be analyzed based on the output power during a certain output time period.

[0036] Step S32: Calculate the distance between each sample data of other samples in the observation data sample and the M cluster center data, and assign each sample data to the class corresponding to the cluster center data with the smallest distance;

[0037] The formula for calculating the distance between sample data and cluster center data is as follows:

[0038]

[0039] Among them, D i,c X represents the distance between sample data i and cluster center data c. i1 X represents the output time period of the i-th sample data. i2 This indicates that the i-th sample data is in X. i1 Charging method during the output period, X i3 This indicates that the i-th sample data is in X. i1 Output power during the output time period, X c1 X represents the output time period of cluster center data c. c2 This indicates that the cluster center data c is in X. c1 Charging method during the output period, X c3 This indicates that the cluster center data c is in X. c1 Output power during the output period.

[0040] In addition, the distance between the sample data and the cluster center data can also be measured using Chebyshev distance, Manhattan distance, etc.

[0041] Step S33: For the M classes obtained, calculate the mean of the samples in each class as the new cluster center data, and determine whether the iteration termination condition is met. If it is met, output the M classes obtained as the cluster structure. If it is not met, repeat steps S32-S33 until the iteration termination condition is met.

[0042] Specifically, the cost function corresponding to the clustering result is calculated, and the above iteration termination condition is satisfied when the cost function corresponding to the clustering result is minimized.

[0043] The cost function corresponding to the clustering results is calculated using the following formula:

[0044]

[0045] Where CF represents the cost function corresponding to the clustering result, i represents the sample data index, T represents the total number of sample data, c represents the cluster center data, M represents the total number of load types, and d ic P represents the distance from sample data i to cluster center data c. ic This indicates the probability that sample data i is clustered into the class corresponding to the cluster center data c.

[0046] Specifically, the above P ic Obtained from the following formula:

[0047]

[0048] Where ρ represents the class boundary adjustment coefficient, d i1 d represents the distance from sample data i to the first cluster center data. i2 d represents the distance from sample data i to the second cluster center data. iM This represents the distance from sample data i to the Mth cluster center data.

[0049] The aforementioned class boundary adjustment coefficient ρ is proportional to the total number of sample data T, i.e., ρ = β × T, where β represents the proportionality coefficient. It should be noted that a larger value of ρ results in clearer boundaries between classes, making any sample data i more likely to be clustered into a single, distant class, leading to clustering errors where data originally located at class boundaries are located, thus reducing clustering accuracy. Conversely, a smaller value of ρ results in more blurred boundaries between classes, making any sample data i more likely to be clustered into more than one class (with the same distance from multiple class centers), but this may cause frequent re-clustering of sample data, increasing the complexity of the clustering process. Therefore, an appropriate value of ρ needs to be set. In this invention, the value of ρ is set to be proportional to T; that is, the larger the total number of sample data, the larger the value of ρ, and the smaller the total number of sample data, the smaller the value of ρ, thus balancing the accuracy and complexity of the clustering process.

[0050] This invention also proposes a cluster charging load spatiotemporal characteristic analysis system for executing the aforementioned cluster charging load spatiotemporal characteristic analysis method. This system includes a data acquisition unit and a data analysis unit. The data acquisition unit is connected to the data analysis unit, and is used to acquire spatial information and charging information of the charging pile cluster. The spatial information includes the location information of each charging pile or sub-group of charging piles within the cluster, and the charging information includes the time-output power curves and charging method information of each charging pile or sub-group of charging piles. The data analysis unit is used to perform spatiotemporal characteristic analysis of the cluster charging load.

[0051] The embodiments and / or implementation methods described above are merely preferred embodiments and / or implementation methods for implementing the technology of the present invention, and are not intended to limit the implementation methods of the technology of the present invention in any way. Any person skilled in the art can make some modifications or alterations to other equivalent embodiments without departing from the scope of the technical means disclosed in the content of the present invention, but they should still be regarded as the technology or embodiments that are substantially the same as the present invention.

[0052] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. The above descriptions are only preferred embodiments of this application. It should be noted that due to the limitations of written expression, while there are objectively infinite specific structures, those skilled in the art can make several improvements, modifications, or changes without departing from the principles of this application, and can also combine the above technical features in an appropriate manner. These improvements, modifications, changes, or combinations, or the direct application of the inventive concept and technical solution to other situations without modification, should all be considered within the scope of protection of this application.

Claims

1. A method for analyzing the spatiotemporal characteristics of cluster charging load, characterized in that, Includes the following steps: Step S1: Obtain the spatial information and charging information of the charging pile cluster; wherein, the spatial information of the charging pile cluster includes the location information of each charging pile or subgroup of charging piles in the charging pile cluster, and the charging information of the charging pile cluster includes the time-output power curve and charging method information of each charging pile or subgroup of charging piles in the charging pile cluster. Step S2: Based on the time-output power curves and charging method information of each charging pile or subgroup of charging piles in the charging pile cluster in Step S1, obtain observation data samples; Step S3: Cluster the observed data samples from Step S2 using an improved mean clustering algorithm based on class boundary adjustment coefficients to obtain the clustering results; the improved mean clustering algorithm includes: Step S31: Based on the load type, randomly select M sample data from the observation data samples obtained in step S2 as cluster center data to obtain M classes; where M represents the total number of load types; Step S32: Calculate the distance between each sample data of other samples in the observation data sample and the M cluster center data, and assign each sample data to the class corresponding to the cluster center data with the smallest distance; Step S33: For the M classes obtained, calculate the mean of the samples in each class as the new cluster center data, and determine whether the iteration termination condition is met. If it is met, output the M classes obtained as the cluster structure. If it is not met, repeat steps S32 and S33 until the iteration termination condition is met. In step S33, the cost function corresponding to the clustering result is calculated, and the iteration termination condition is satisfied when the cost function corresponding to the clustering result is minimized. The cost function corresponding to the clustering results is calculated using the following formula: ； Where CF represents the cost function corresponding to the clustering result, i represents the sample data index, T represents the total number of sample data, c represents the cluster center data, M represents the total number of load types, and d ic P represents the distance from sample data i to cluster center data c. ic This indicates the probability that sample data i is clustered into the class corresponding to cluster center data c; The P ic Obtained from the following formula: ； Where ρ represents the class boundary adjustment coefficient, d i1 d represents the distance from sample data i to the first cluster center data. i2 d represents the distance from sample data i to the second cluster center data. iM This represents the distance from sample data i to the Mth cluster center data. The class boundary adjustment coefficient ρ is proportional to the total number of sample data T, i.e., ρ = β × T, where β represents the proportionality coefficient; Step S4: Based on the location information of each charging pile or subgroup of charging piles in Step S1 and the clustering results in Step S3, perform spatiotemporal characteristic analysis of cluster charging load.

2. The method for analyzing the spatiotemporal characteristics of cluster charging load according to claim 1, characterized in that, In step S2, the charging method information includes slow charging and fast charging.

3. The method for analyzing the spatiotemporal characteristics of cluster charging load according to claim 1, characterized in that, In step S2, the observation data sample includes multiple sample data Di, where i represents the sample data sequence number, and Di = [X]. i1 X i2 X i3 ], where X i1 X represents the output time period of the i-th sample data. i2 This indicates that the i-th sample data is in X. i1 Charging method during the output period, X i3 This indicates that the i-th sample data is in X. i1 Output power during the output period.

4. The method for analyzing the spatiotemporal characteristics of cluster charging load according to claim 1, characterized in that, In step S32, the formula for calculating the distance between the sample data and the cluster center data is: ； Among them, D i,c X represents the distance between sample data i and cluster center data c. i1 X represents the output time period of the i-th sample data. i2 This indicates that the i-th sample data is in X. i1 Charging method during the output period, X i3 This indicates that the i-th sample data is in X. i1 Output power during the output time period, X c1 X represents the output time period of cluster center data c. c2 This indicates that the cluster center data c is in X. c1 Charging method during the output period, X c3 This indicates that the cluster center data c is in X. c1 Output power during the output period.

5. A system for analyzing the spatiotemporal characteristics of cluster charging load, used to execute the method for analyzing the spatiotemporal characteristics of cluster charging load as described in any one of claims 1-4, characterized in that, It includes a data acquisition unit and a data analysis unit, the data acquisition unit being connected to the data analysis unit. The data acquisition unit is used to acquire spatial information of the charging pile cluster and charging information of the charging pile cluster. The spatial information of the charging pile cluster includes the location information of each charging pile or subgroup of charging piles in the charging pile cluster, and the charging information of the charging pile cluster includes the time-output power curve and charging method information of each charging pile or subgroup of charging piles in the charging pile cluster. The data analysis unit is used to perform spatiotemporal characteristic analysis of the cluster charging load.

Images