Multi-charging station user load prediction method and device, and power selling agent system

Abstract

The application discloses a multi-charging station user load prediction method and device and a power selling agent system, and the method comprises the following steps: acquiring the position information, historical order information and historical time-of-use electricity information of a plurality of charging stations; constructing a flow complementary adjacency matrix according to the position information and historical order information of each charging station; aggregating the plurality of charging stations into at least one station cluster according to the flow complementary adjacency matrix; predicting the load component of each station cluster according to the historical time-of-use electricity information of each charging station based on a trained load prediction regression model; and determining the total load of the plurality of charging stations according to the load component of each station cluster. The application makes a more accurate description of the user portrait, greatly improves the prediction accuracy, has simple algorithm logic, low computing power requirement, significantly reduces the cost of charging station load prediction, has strong practicability, and provides accurate reference value for power purchase of the power selling agent company.

Description

Technical Field

[0001] This application relates to the field of power data technology, specifically to a method, device, and electricity sales agency system for predicting user load at multiple charging stations. Background Technology

[0002] In the electricity market, electricity retailers act as agents, purchasing electricity on behalf of users. These retailers need to accurately predict the total electricity load of users in order to profit from this business.

[0003] The users of electricity sales companies of charging service providers or charging operators are mostly users of charging stations, and most existing load forecasting algorithms cannot accurately depict user profiles, thus reducing the accuracy of power load forecasting. Summary of the Invention

[0004] In response to the above-mentioned situation, this application proposes a method, device, and electricity sales agency system for predicting the user load of multiple charging stations, in order to overcome or at least partially overcome the shortcomings of the prior art.

[0005] In a first aspect, embodiments of this application provide a method for predicting user load at multiple charging stations, including:

[0006] Obtain location information, historical order information, and historical time-of-use electricity consumption information for multiple charging stations;

[0007] Based on the location information and historical order information of each charging station, a traffic complementary adjacency matrix is ​​constructed;

[0008] Based on the traffic complementary adjacency matrix, the multiple charging stations are aggregated into at least one station cluster;

[0009] Based on the trained load prediction regression model, the load components of each charging station cluster are predicted according to the historical time-of-use electricity consumption information of each station.

[0010] The total load of the multiple charging stations is determined based on the load components of each station cluster.

[0011] Optionally, in the above method, the step of constructing a traffic complementary adjacency matrix based on the location information and historical order information of each charging station includes:

[0012] Based on the location information of each charging station and the pre-built geographical complementarity model, the geographical proximity complementarity of each pair of charging stations is determined.

[0013] Based on the historical order information of each charging station and the pre-built habit complementarity model, the charging habit complementarity of each pair of charging stations is determined;

[0014] Based on preset weight parameters and the complementarity of geographical proximity and charging habits between each pair of charging stations, the traffic complementarity between each pair of charging stations is determined.

[0015] Initialize the flow complementary adjacency matrix and fill the corresponding positions of the flow complementary adjacency matrix with the flow complementarity of every two charging stations.

[0016] Optionally, in the above method, the geographical complementarity model is:

[0017] L(Xi,Xj)=1-min(D(Xi,Xj) / α,1);

[0018] Where D is the latitude and longitude distance between charging station Xi and charging station Xj, and α is a preset distance threshold.

[0019] The habit complementarity model is as follows:

[0020]

[0021] Where Intersection(Xi, Xj) represents the number of orders for the same vehicle ID at charging stations Xi and Xj, and O(Xi) and O(Xj) represent the number of orders at charging stations Xi and Xj, respectively.

[0022] Optionally, in the above method, the step of aggregating the multiple charging stations into at least one station cluster based on the traffic complementary adjacency matrix includes:

[0023] Determine the degree matrix based on the aforementioned complementary adjacency matrix;

[0024] Based on the degree matrix, the flow complementary adjacency matrix is ​​transformed and standardized sequentially to obtain the Laplace matrix;

[0025] Based on the Laplacian matrix, a feature matrix is ​​determined, where each row of the feature matrix represents an eigenvalue of the Laplacian matrix and the corresponding eigenvector.

[0026] The feature matrix is ​​standardized row by row to obtain multiple samples to be clustered, and each sample to be clustered corresponds to a charging station.

[0027] Clustering multiple samples to be clustered yields at least one site cluster.

[0028] Optionally, in the above method, the step of predicting the load components of each charging station cluster based on the trained load prediction regression model and the historical time-of-use electricity consumption information of each charging station includes:

[0029] The historical time-of-use electricity consumption information of each charging station in each of the aforementioned station clusters is accumulated to obtain the historical electricity consumption time series data of each of the aforementioned station clusters;

[0030] The historical electricity consumption time series data of each of the aforementioned stations are input into the load prediction regression model to extract electricity consumption features, which include: time-of-use electricity consumption, time, temperature, and electricity consumption during holidays;

[0031] Load inference is performed based on the power consumption characteristics of each of the aforementioned power station clusters to obtain the load components of each of the aforementioned power station clusters within a preset future time period.

[0032] Optionally, in the above method, determining the total load of the multiple charging stations based on the load components of various types of charging stations includes:

[0033] The load components of various charging stations are summed to obtain the total load of the multiple charging stations.

[0034] Secondly, embodiments of this application also provide a multi-charging station user load prediction device, the device comprising:

[0035] The acquisition unit is used to acquire location information, historical order information, and historical time-of-use electricity consumption information of multiple charging stations;

[0036] The construction unit is used to construct a traffic complementary adjacency matrix based on the location information and historical order information of each charging station;

[0037] An aggregation unit is used to aggregate the plurality of charging stations into at least one station cluster according to the traffic complementary adjacency matrix.

[0038] The prediction unit is used to predict the load components of each charging station cluster based on the historical time-of-use electricity consumption information of each station, according to the trained load prediction regression model.

[0039] The aggregation unit is used to determine the total load of the multiple charging stations based on the load components of each station cluster.

[0040] Thirdly, this application also provides an electricity sales agency system, which includes interconnected electricity trading devices and the aforementioned multi-charging station user load forecasting device. The electricity sales agency system can also be connected to the electricity market system and multiple charging stations.

[0041] The power trading device is used to purchase electricity from the power market system based on the total load obtained by the multi-charging station user load forecasting device, and then sell it to each charging station.

[0042] Fourthly, embodiments of this application also provide an electronic device, including: a processor; and a memory arranged to store computer-executable instructions, which, when executed, cause the processor to perform any of the methods described above.

[0043] Fifthly, embodiments of this application also provide a computer-readable storage medium that stores one or more programs, which, when executed by an electronic device including multiple applications, cause the electronic device to perform any of the methods described above.

[0044] The above-described technical solutions adopted in the embodiments of this application can achieve the following beneficial effects:

[0045] This application incorporates the complementary nature of charging station traffic flow into the prediction of charging station electricity load. By using the location information and historical order information of each charging station, a traffic complementarity adjacency matrix is ​​constructed. The elements of the traffic complementarity adjacency matrix represent the traffic complementarity between two charging stations. Multiple charging stations are aggregated into one or more complementary categories through the traffic complementarity adjacency matrix, forming station clusters. Then, multiple charging stations in each station cluster are input into a trained load prediction regression model to obtain the load component of each station cluster. Finally, the load components of all station clusters are summed to obtain the total load of multiple charging stations. This application constructs an adjacency matrix based on the complementary characteristics of traffic flow among charging stations. Using this adjacency matrix and a spectral clustering algorithm, the stations are clustered. A trained load prediction regression model is then used to predict the total traffic flow of charging stations within the same cluster. These predictions are then combined to obtain the final total load prediction result. This application provides a more accurate user profile, significantly improving prediction accuracy. Furthermore, the algorithm is simple in logic, requires low computational power, and significantly reduces the cost of charging station load prediction. It is highly practical and provides accurate reference value for electricity purchases by electricity sales agents. Attached Figure Description

[0046] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0047] Figure 1 A schematic diagram of an application scenario according to this application is shown;

[0048] Figure 2 A flowchart illustrating a multi-charging station user load prediction method according to an embodiment of this application is shown;

[0049] Figure 3A schematic diagram of the structure of a traffic complementary adjacency matrix according to an embodiment of this application is shown;

[0050] Figure 4 A flowchart illustrating a multi-charging station user load prediction method according to another embodiment of this application is shown;

[0051] Figure 5 A flowchart illustrating a multi-charging station user load forecasting method according to another embodiment of this application is shown.

[0052] Figure 6 A schematic diagram of the structure of a multi-charging station user load prediction device according to an embodiment of this application is shown;

[0053] Figure 7 A schematic diagram of the structure of an electricity sales agency system according to an embodiment of this application is shown;

[0054] Figure 8 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0056] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0057] The concept of this application is to introduce the characteristics of traffic complementarity between charging stations into user profiling, construct an adjacency matrix representing the traffic complementarity between charging stations, and cluster the stations based on the adjacency matrix and spectral clustering algorithm to obtain one or more station clusters. The total traffic of each charging station in the same station cluster is predicted, and finally the results are superimposed to obtain the total load prediction result, so as to improve the prediction accuracy.

[0058] Figure 1 A schematic diagram of an application scenario according to this application is shown, from Figure 1 As can be seen, in the electricity market, the electricity sales company acts as an agent to purchase electricity from the electricity market on behalf of electricity users. The overall process can be roughly divided into three stages. First, load forecasting; then, purchasing electricity from the electricity market based on the load forecast results; and finally, selling it to various charging stations to complete the transaction. In this transaction process, the electricity sales company needs to accurately predict the total electricity load of the electricity users in order to profit from the electricity purchase business.

[0059] In existing technologies, the accuracy of user load prediction is relatively low. To address this, this application provides a method for predicting user load at multiple charging stations. Figure 2 A flowchart illustrating a multi-charging station user load forecasting method according to an embodiment of this application is shown. Figure 2 It can be seen that this application includes at least steps S210 to S250:

[0060] Step S210: Obtain the location information, historical order information, and historical time-of-use electricity consumption information of multiple charging stations.

[0061] When forecasting the load of multiple charging stations, the first step is to obtain the data information required for the forecasting process. Taking a single charging station as an example, the required data includes, but is not limited to: location information, historical order information, and historical time-of-use electricity information.

[0062] Location information refers to the geographical location of a charging station. In this application, latitude and longitude data can be used to represent the location information of the charging station.

[0063] Historical order information refers to order information generated by a charging station within a certain historical period (such as the past year or a quarter). Order information may include, but is not limited to, vehicle ID (such as license plate number), vehicle charging time, charging amount, payment amount, etc. Historical order information is collected with user authorization.

[0064] Historical time-of-use electricity consumption information refers to the electricity consumption of a charging station at different times within a historical period (such as the past year or a quarter). Specifically, it can be collected through the IoT data of the main electricity meter of each charging station or the SaaS (a management system) system of the charging pile underlying platform.

[0065] Step S220: Construct a traffic complementary adjacency matrix based on the location information and historical order information of each charging station.

[0066] After obtaining the data required for load forecasting, a traffic complementary adjacency matrix is ​​constructed based on the location information and historical order information of each charging station.

[0067] Different charging stations have complementary flow characteristics in the following ways:

[0068] 1. Complementarity of neighboring charging stations: Electric vehicle owners will choose to charge at one of several geographically close charging stations, which makes traffic complementarity between neighboring charging stations, that is, the fluctuation of their total traffic is less than the fluctuation of their individual traffic.

[0069] 2. Complementary charging habits: Electric vehicle owners are accustomed to charging at a few relatively fixed charging stations due to their daily work and life trajectories. This makes the traffic flow between these charging stations complementary, that is, the fluctuation of their total traffic flow is less than the fluctuation of their individual traffic flow.

[0070] Based on the complementarity of neighboring charging stations and the complementarity of charging habits, an adjacency matrix can be constructed to represent the traffic complementarity among multiple charging stations. This matrix is ​​denoted as the traffic complementarity adjacency matrix. The traffic complementarity adjacency matrix can be a square matrix, and its shape can be as follows: Figure 3 As shown, Figure 3 A schematic diagram of the structure of a traffic complementary adjacency matrix according to an embodiment of this application is shown. Figure 3 As can be seen, the diagonal of the flow complementary adjacency matrix represents the intersection of each charging station with itself, with each element having a value of 1 (or other preset values). The off-diagonal of the flow complementary adjacency matrix represents the intersection of one charging station with another, with each element representing the flow complementarity between the two charging stations. It should be noted that the corresponding elements of the off-diagonal lines, S12 = S21, S32 = S23, and so on, are derived accordingly.

[0071] In some embodiments of this application, traffic complementarity mainly includes two aspects: geographical proximity complementarity and charging habit complementarity. The step of constructing a traffic complementarity adjacency matrix based on the location information and historical order information of each charging station includes: determining the geographical proximity complementarity of every two charging stations based on the location information of each charging station and a pre-built geographical complementarity model; determining the charging habit complementarity of every two charging stations based on the historical order information of each charging station and a pre-built habit complementarity model; determining the traffic complementarity of every two charging stations based on preset weight parameters and the geographical proximity complementarity and charging habit complementarity of every two charging stations; initializing the traffic complementarity adjacency matrix and filling the corresponding positions of the traffic complementarity adjacency matrix with the traffic complementarity of every two charging stations.

[0072] In some embodiments of this application, traffic complementarity can be the sum of geographical proximity complementarity and charging habit complementarity. In other embodiments, a weight parameter can be assigned to both, and the relative weights of geographical proximity complementarity and charging habit complementarity can be set by setting this weight parameter. Please refer to... Figure 3 ,for Figure 3 For any off-diagonal element, its flow complementarity S ij It can be expressed using equation (1):

[0073] S ij =L(Xi,Xj)+βH(Xi,Xj) Formula (1);

[0074] L(Xi,Xj) represents the geographical proximity complementarity between charging stations Xi and Xj, and H(Xi,Xj) represents the charging habit complementarity between charging stations Xi and Xj; β represents a predefined weight parameter; by setting β, the relative weights of the two items, geographical proximity complementarity and charging habit complementarity, can be set.

[0075] Regarding the geographical proximity complementarity between power station Xi and charging station Xj, the geographical proximity complementarity between each pair of charging stations can be determined based on the location information of power station Xi and charging station Xj and a pre-constructed geographical complementarity model. In some embodiments of this application, the geographical complementarity model can be expressed as Equation (2):

[0076] L(Xi,Xj)=1-min(D(Xi,Xj) / α,1) Formula (2);

[0077] Where D is the latitude and longitude distance between charging station Xi and charging station Xj, and its calculation formula is shown in equation (3):

[0078] D(A,B)=111.12cos{1 / [sinΦAsinΦB+cosΦAcosΦBcos(λB—λA)]} Formula (3);

[0079] Where A represents the location of charging station Xi, B represents the location of charging station Xj, the longitude and latitude of point A are λA and ΦA respectively, and the longitude and latitude of point B are λB and ΦB respectively;

[0080] α is a predefined distance threshold. Two charging stations that are more than α apart are considered to have no geographical proximity complementarity.

[0081] As can be seen from equation (2), by inputting the latitude and longitude information of charging station Xi and charging station Xj into equation (2), the geographical proximity complementarity between charging station Xi and charging station Xj can be determined.

[0082] Regarding the complementarity of charging habits between power station Xi and charging station Xj, the complementarity of charging habits between power station Xi and charging station Xj can be determined based on the location information of power station Xi and charging station Xj, as well as the pre-built historical order information and the pre-built habit complementarity model. In some embodiments of this application, the charging habit complementarity model can be expressed as Equation (4):

[0083]

[0084] Where Intersection(Xi, Xj) represents the number of orders for the same vehicle ID at charging stations Xi and Xj, and O(Xi) and O(Xj) represent the number of orders at charging stations Xi and Xj, respectively.

[0085] As can be seen from equation (4), by inputting the historical order information of charging station Xi and charging station Xj into equation (4), the complementarity of charging habits between charging station Xi and charging station Xj can be determined.

[0086] After determining the flow complementarity S between every two charging stations ij Then, fill the corresponding positions in the initialized flow complementary adjacency matrix to obtain the flow complementary adjacency matrix of multiple charging stations. Here, the flow complementary adjacency matrix is ​​denoted as matrix S.

[0087] Step S230: Aggregate the multiple charging stations into at least one station cluster according to the traffic complementary adjacency matrix.

[0088] After obtaining the traffic complementary adjacency matrix S, all charging stations can be aggregated into one or more clusters based on this matrix. In practical scenarios, multiple clusters are usually used, and each cluster is denoted as a station cluster. Each station cluster contains one or more charging stations, usually more. For example, by processing the traffic complementary adjacency matrix S, the multiple charging stations can be divided into k2 clusters based on the N-Cut spectral clustering algorithm.

[0089] Specifically, in some embodiments of this application, the step of aggregating the multiple charging stations into at least one station cluster based on the traffic complementary adjacency matrix includes: determining a degree matrix based on the traffic complementary adjacency matrix; transforming and standardizing the traffic complementary adjacency matrix sequentially based on the degree matrix to obtain a Laplace matrix; determining a feature matrix based on the Laplace matrix, where each row of the feature matrix represents an eigenvalue of the Laplace matrix and the eigenvector corresponding to the eigenvalue; standardizing the feature matrix row by row to obtain multiple samples to be clustered, each of the samples to be clustered corresponding to a charging station; and clustering the multiple samples to be clustered to obtain at least one station cluster.

[0090] Assuming there are n charging stations, firstly, based on the flow complementary adjacency matrix S, the degree matrix D is determined. The degree matrix D is a 1×n vector, where each element can be represented by equation (5):

[0091]

[0092] Then, based on the flow complementary adjacency matrix S and the degree matrix D, the Laplace matrix L is calculated. The Laplace matrix L = DS. The calculated Laplace matrix L is unnormalized and needs to be normalized. The normalized Laplace matrix is: D -1 / LD -1 / The specific standardization process can be found in existing technologies, and will not be elaborated here.

[0093] Then calculate the normalized Laplace matrix D. -1 / LD -1 / This yields k1 eigenvalues ​​and k1 corresponding eigenvectors f. It should be noted that the k1 eigenvalues ​​can be selected from the standardized Laplace matrix D, depending on the specific needs. -1 / 2 LD -1 / Alternatively, all feature values ​​can be arranged in ascending order, and then the first k1 feature values ​​can be extracted.

[0094] Each feature vector f is treated as a row. The resulting k1 feature vectors f are arranged in ascending order of their k1 eigenvalues ​​to form a feature matrix, denoted as feature matrix F. Feature matrix F is then standardized row by row to form an n×k1 dimensional standardized feature matrix F. Each row of the standardized feature matrix F is taken as a k1 dimensional sample to be clustered, resulting in n k1 dimensional samples to be clustered. The feature information contained in each sample to be clustered corresponds to a charging station.

[0095] For n samples to be clustered, a clustering method is used to perform clustering with a clustering dimension of k2, resulting in multiple station clusters, which are denoted as c1, c2, ... ck2.

[0096] Step S240: Based on the trained load prediction regression model, predict the load components of each charging station cluster according to the historical time-of-use electricity consumption information of each charging station.

[0097] After clustering multiple charging stations, the load forecasting stage begins. In this application, during pre-testing, forecasts are made for each type of charging station cluster. Please refer to [link / reference]. Figure 4 , Figure 4 A flowchart illustrating a multi-charging station user load forecasting method according to another embodiment of this application is shown. Figure 4 As can be seen, firstly, many charging stations are clustered to obtain station cluster A, station cluster B, and station cluster C. Then, load prediction is performed on the stations in each cluster, and the total load of station cluster A, station cluster B, and station cluster C are obtained. Finally, the loads of each station cluster are summed to obtain the total load of all charging stations.

[0098] In this application, the load of each charging station cluster is referred to as a load component. The following is a simple explanation of the prediction process using the prediction of the load component of a single charging station cluster as an example. Load prediction can be achieved using a pre-trained load prediction regression model. This model can be constructed using any existing network or algorithm, such as the XGBoost algorithm. A large amount of data is used for training. The training process can be summarized as follows: a large amount of training data is collected, pre-processed to form a training dataset, and each data point in the training dataset is input into the load prediction regression model in segments or batches based on the historical electricity consumption time-series data of a charging station. Feature extraction is performed, including but not limited to: total electricity consumption time-series data, year, month, day, time, temperature, and holidays. The loss function of the load prediction regression model minimizes the mean square error of the prediction. The parameters of the load prediction regression model are backpropagated based on the value of the loss function at each step. Once the training requirements are met, the training ends, and the pre-trained load prediction regression model is obtained.

[0099] The logic of the load forecasting process is consistent with that of the load forecasting regression model training process. Specifically, in some embodiments of this application, based on the trained load forecasting regression model, the load components of each charging station cluster are predicted according to the historical time-of-use electricity consumption information of each charging station. This includes: accumulating the historical time-of-use electricity consumption information of each charging station in each charging station cluster to obtain the historical electricity consumption time series data of each charging station cluster; inputting the historical electricity consumption time series data of each charging station cluster into the load forecasting regression model for electricity consumption feature extraction, the features including: time-of-use electricity consumption, time, temperature, and holiday electricity consumption; and performing load inference based on the electricity consumption features of each charging station cluster to obtain the load components of each charging station cluster within a preset future time period.

[0100] In other words, for a charging station cluster, the input data is the historical time-of-use electricity consumption information of all charging stations in the cluster, and the output data is the load forecast result (load component) of the cluster in the future (such as the next year, the next quarter, or the next month).

[0101] Specifically, the historical time-of-use electricity consumption information of all charging stations in the cluster is aggregated and processed to form the historical electricity consumption time-series data for that cluster. This data is then input into a load forecasting regression model. The model extracts features from the historical electricity consumption time-series data, consistent with the training process. Extracted features include, but are not limited to, time-of-use electricity consumption, time, temperature, and electricity consumption during holidays. Based on these extracted features, inferences and predictions are made to obtain the load components for the cluster within a predetermined future timeframe.

[0102] For each station cluster, the load forecasting model described above is used to infer the load components for each station cluster over a future period. Similar to the previous embodiment, k2 station clusters generate a total of k2 load forecast results, i.e., k2 load components.

[0103] Step S250: Determine the total load of the multiple charging stations based on the load components of each station cluster.

[0104] By summing up each load component, the total load of all charging stations can be obtained. This involves summing up k2 prediction results to get the predicted total load of all charging stations. Subsequently, the electricity sales company uses this total load to purchase electricity on behalf of users in the electricity market.

[0105] Depend on Figure 2 As can be seen, this application incorporates the complementary nature of charging station traffic into the electricity load prediction of charging stations. By using the location information and historical order information of each charging station, a traffic complementarity adjacency matrix is ​​constructed. The elements of the traffic complementarity adjacency matrix represent the traffic complementarity of two charging stations. Multiple charging stations are aggregated into one or more complementary categories through the traffic complementarity adjacency matrix, forming station clusters. Then, multiple charging stations in each station cluster are input into a trained load prediction regression model to obtain the load component of each station cluster. Finally, the load components of all station clusters are summed to obtain the total load of multiple charging stations. This application constructs an adjacency matrix based on the complementary characteristics of traffic flow among charging stations. Using this adjacency matrix and a spectral clustering algorithm, the stations are clustered. A trained load prediction regression model is then used to predict the total traffic flow of charging stations within the same cluster. These predictions are then combined to obtain the final total load prediction result. This application provides a more accurate user profile, significantly improving prediction accuracy. Furthermore, the algorithm is simple in logic, requires low computational power, and significantly reduces the cost of charging station load prediction. It is highly practical and provides accurate reference value for electricity purchases by electricity sales agents.

[0106] Figure 5 A flowchart illustrating a multi-charging station user load forecasting method according to another embodiment of this application is shown. Figure 5 As can be seen, this embodiment includes steps S510 to S570:

[0107] Step S510: Obtain the total number n of all charging stations, as well as the location information, historical order information, and historical time-of-use electricity consumption information of each charging station.

[0108] Step S520: Initialize the flow complementary adjacency matrix according to the total number n of all charging stations.

[0109] Step S530: Based on the location information and historical order information of each charging station, determine the traffic complementarity between every two charging stations, and place them in the corresponding positions in the initialized traffic complementarity adjacency matrix to obtain the traffic complementarity adjacency matrix.

[0110] Step S540: Aggregate the n charging stations according to the flow complementary adjacency moments to obtain m types of station clusters, where m is less than n.

[0111] Step S550: Process the historical time-of-use electricity consumption information of each charging station in the m-type station cluster to obtain m historical electricity consumption time series data.

[0112] Step S560: Input the m historical electricity consumption time series data into the load prediction regression model to obtain the load components of the m types of power station clusters.

[0113] Step S570: Sum the m load components to obtain the total load of all charging stations.

[0114] Figure 6 A schematic diagram of a multi-charging station user load prediction device according to an embodiment of this application is shown. Figure 6 It can be seen that the multi-charging station user load prediction device 600 includes:

[0115] The acquisition unit 610 is used to acquire location information, historical order information, and historical time-of-use electricity consumption information of multiple charging stations;

[0116] The construction unit 620 is used to construct a traffic complementary adjacency matrix based on the location information and historical order information of each charging station;

[0117] Aggregation unit 630 is used to aggregate the plurality of charging stations into at least one station cluster according to the traffic complementary adjacency matrix;

[0118] The prediction unit 640 is used to predict the load components of each charging station cluster based on the historical time-of-use electricity consumption information of each charging station, according to the trained load prediction regression model.

[0119] The aggregation unit 650 is used to determine the total load of the plurality of charging stations based on the load components of each of the station clusters.

[0120] In some embodiments of this application, in the above-described apparatus, the construction unit 620 is configured to: determine the geographical proximity complementarity of each pair of charging stations based on the location information of each charging station and a pre-built geographical complementarity model; determine the charging habit complementarity of each pair of charging stations based on the historical order information of each charging station and a pre-built habit complementarity model; determine the traffic complementarity of each pair of charging stations based on preset weight parameters and the geographical proximity complementarity and charging habit complementarity of each pair of charging stations; initialize the traffic complementarity adjacency matrix; and fill the corresponding positions of the traffic complementarity of each pair of charging stations into the traffic complementarity adjacency matrix.

[0121] In some embodiments of this application, in the above-described apparatus, the geographical complementarity model is:

[0122] L(Xi,Xj)=1-min(D(Xi,Xj) / α,1);

[0123] Where D is the latitude and longitude distance between charging station Xi and charging station Xj, and α is a preset distance threshold.

[0124] The habit complementarity model is as follows:

[0125]

[0126] Where Intersection(Xi, Xj) represents the number of orders for the same vehicle ID at charging stations Xi and Xj, and O(Xi) and O(Xj) represent the number of orders at charging stations Xi and Xj, respectively.

[0127] In some embodiments of this application, in the above-described apparatus, the aggregation unit 630 is configured to: determine a degree matrix based on the traffic complementary adjacency matrix; transform and standardize the traffic complementary adjacency matrix sequentially based on the degree matrix to obtain a Laplace matrix; determine a feature matrix based on the Laplace matrix, wherein each row of the feature matrix represents an eigenvalue of the Laplace matrix and the eigenvector corresponding to the eigenvalue; standardize the feature matrix row by row to obtain multiple samples to be clustered, each of the samples to be clustered corresponding to a charging station; and cluster the multiple samples to be clustered to obtain at least one charging station cluster.

[0128] In some embodiments of this application, in the above-described apparatus, the prediction unit 640 is used to accumulate the historical time-of-use electricity consumption information of each charging station in each of the said station clusters to obtain the historical electricity consumption time series data of each of the said station clusters; input the historical electricity consumption time series data of each of the said station clusters into the load prediction regression model to extract electricity consumption features, the features including: time-of-use electricity consumption, time, temperature, and electricity consumption during holidays; and perform load inference based on the electricity consumption features of each of the said station clusters to obtain the load component of each of the said station clusters within a preset future time period.

[0129] In some embodiments of this application, in the above-described apparatus, the summing unit 650 is used to sum the load components of various charging stations to obtain the total load of the plurality of charging stations.

[0130] It should be noted that the aforementioned multi-charging station user load prediction device can implement the aforementioned multi-charging station user load prediction method one by one, and will not be described again.

[0131] Figure 7 A schematic diagram of a power sales agency system according to an embodiment of this application is shown. The power sales agency system 700 can be deployed in a power sales company and... Figure 7 As can be seen, the electricity sales agency system 700 includes interconnected electricity trading devices 710 and the multi-charging station user load forecasting device 600 described in any of the preceding items. The electricity sales agency system 700 can be connected to the electricity market system 800 and multiple charging stations 900. The electricity trading device 710 purchases electricity from the electricity market system 800 based on the total load obtained by the multi-charging station user load forecasting device 600 and sells it to each charging station 900.

[0132] Figure 8 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Please refer to it. Figure 8 At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and memory. The memory may include main memory, such as high-speed random-access memory (RAM), or non-volatile memory, such as at least one disk drive. Of course, the electronic device may also include other hardware required for other business operations.

[0133] The processor, network interface, and memory can be interconnected via an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 8 The symbol is represented by a single double-headed arrow, but this does not mean that there is only one bus or one type of bus.

[0134] Memory is used to store programs. Specifically, programs may include program code, which includes computer operation instructions. Memory may include main memory and non-volatile memory, and provides instructions and data to the processor.

[0135] The processor reads the corresponding computer program from non-volatile memory into main memory and then runs it, forming a multi-charging station user load forecasting device at the logical level. The processor executes the program stored in memory and specifically performs the aforementioned methods.

[0136] The above is as stated in this application. Figure 6The method executed by the multi-charging station user load forecasting device disclosed in the illustrated embodiment can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in the embodiments of this application can be directly embodied in the execution of a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software module can reside in a mature storage medium in the field, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0137] The electronic device can also perform Figure 6 A method for executing a multi-charging station user load forecasting device, and implementing the multi-charging station user load forecasting device in... Figure 6 The functions of the embodiments shown are not described in detail here.

[0138] This application also proposes a computer-readable storage medium that stores one or more programs, the programs including instructions that, when executed by an electronic device including multiple applications, enable the electronic device to perform... Figure 6 The method executed by the multi-charging station user load prediction device in the illustrated embodiment is specifically used to execute the aforementioned method.

[0139] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0140] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0141] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0142] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0143] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0144] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0145] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0146] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0147] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0148] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for predicting user load at multiple charging stations, characterized in that, include: Obtain location information, historical order information, and historical time-of-use electricity consumption information for multiple charging stations; Based on the location information and historical order information of each charging station, a traffic complementary adjacency matrix is ​​constructed; Based on the traffic complementary adjacency matrix, the multiple charging stations are aggregated into at least one station cluster; Based on the trained load prediction regression model, the load components of each charging station cluster are predicted according to the historical time-of-use electricity consumption information of each station. The total load of the multiple charging stations is determined based on the load components of each station cluster. The step of constructing a traffic complementary adjacency matrix based on the location information and historical order information of each charging station includes: Based on the location information of each charging station and the pre-built geographical complementarity model, the geographical proximity complementarity of each pair of charging stations is determined. Based on the historical order information of each charging station and the pre-built habit complementarity model, the charging habit complementarity of each pair of charging stations is determined; Based on preset weight parameters and the complementarity of geographical proximity and charging habits between each pair of charging stations, the traffic complementarity between each pair of charging stations is determined. Initialize the flow complementary adjacency matrix and fill the corresponding positions of the flow complementary adjacency matrix with the flow complementarity of every two charging stations.

2. The method according to claim 1, characterized in that, The geographical complementarity model is as follows: ； Where D represents the latitude and longitude distance between charging station Xi and charging station Xj. It is a preset distance threshold; The habit complementarity model is as follows: ； in, This indicates the number of orders for the same vehicle ID at charging station Xi and charging station Xj. and These represent the order quantities for charging station Xi and charging station Xj, respectively.

3. The method according to claim 1, characterized in that, The step of aggregating the multiple charging stations into at least one station cluster based on the traffic complementary adjacency matrix includes: Determine the degree matrix based on the aforementioned complementary adjacency matrix; Based on the degree matrix, the flow complementary adjacency matrix is ​​transformed and standardized sequentially to obtain the Laplace matrix; Based on the Laplacian matrix, a feature matrix is ​​determined, where each row of the feature matrix represents an eigenvalue of the Laplacian matrix and the corresponding eigenvector. The feature matrix is ​​standardized row by row to obtain multiple samples to be clustered, and each sample to be clustered corresponds to a charging station. Clustering multiple samples to be clustered yields at least one site cluster.

4. The method according to claim 1, characterized in that, The pre-trained load prediction regression model predicts the load components of each charging station cluster based on their historical time-of-use electricity consumption information, including: The historical time-of-use electricity consumption information of each charging station in each of the aforementioned station clusters is accumulated to obtain the historical electricity consumption time series data of each of the aforementioned station clusters; The historical electricity consumption time series data of each of the aforementioned stations are input into the load prediction regression model to extract electricity consumption features, which include: time-of-use electricity consumption, time, temperature, and electricity consumption during holidays; Load inference is performed based on the power consumption characteristics of each of the aforementioned power station clusters to obtain the load components of each of the aforementioned power station clusters within a preset future time period.

5. The method according to claim 1, characterized in that, Determining the total load of the multiple charging stations based on the load components of various types of charging stations includes: The load components of various charging stations are summed to obtain the total load of the multiple charging stations.

6. A multi-charging station user load prediction device, characterized in that, The device includes: The acquisition unit is used to acquire location information, historical order information, and historical time-of-use electricity consumption information of multiple charging stations; The construction unit is used to construct a traffic complementary adjacency matrix based on the location information and historical order information of each charging station; An aggregation unit is used to aggregate the plurality of charging stations into at least one station cluster according to the traffic complementary adjacency matrix. The prediction unit is used to predict the load components of each charging station cluster based on the historical time-of-use electricity consumption information of each station, according to the trained load prediction regression model. The aggregation unit is used to determine the total load of the multiple charging stations based on the load components of each station cluster. The construction unit is used to determine the geographical proximity complementarity of every two charging stations based on the location information of each charging station and a pre-built geographical complementarity model. Based on the historical order information of each charging station and the pre-built habit complementarity model, the charging habit complementarity of each pair of charging stations is determined; Based on preset weight parameters and the complementarity of geographical proximity and charging habits between each pair of charging stations, the traffic complementarity between each pair of charging stations is determined. Initialize the flow complementary adjacency matrix and fill the corresponding positions of the flow complementary adjacency matrix with the flow complementarity of every two charging stations.

7. The apparatus according to claim 6, characterized in that, The geographical complementarity model is as follows: ； Where D represents the latitude and longitude distance between charging station Xi and charging station Xj. It is a preset distance threshold; The habit complementarity model is as follows: ； in, This indicates the number of orders for the same vehicle ID at charging station Xi and charging station Xj. and These represent the order quantities for charging station Xi and charging station Xj, respectively.

8. The apparatus according to claim 6, characterized in that, The aggregation unit is used to determine the degree matrix based on the traffic complementary adjacency matrix; Based on the degree matrix, the flow complementary adjacency matrix is ​​transformed and standardized sequentially to obtain the Laplace matrix; Based on the Laplacian matrix, a feature matrix is ​​determined, where each row of the feature matrix represents an eigenvalue of the Laplacian matrix and the corresponding eigenvector. The feature matrix is ​​standardized row by row to obtain multiple samples to be clustered, and each sample to be clustered corresponds to a charging station. Clustering multiple samples to be clustered yields at least one site cluster.

9. The apparatus according to claim 6, characterized in that, The prediction unit is used to accumulate the historical time-of-use electricity consumption information of each charging station in each of the station clusters to obtain the historical electricity consumption time series data of each of the station clusters. The historical electricity consumption time series data of each of the aforementioned stations are input into the load prediction regression model to extract electricity consumption features, which include: time-of-use electricity consumption, time, temperature, and electricity consumption during holidays; Load inference is performed based on the power consumption characteristics of each of the aforementioned power station clusters to obtain the load components of each of the aforementioned power station clusters within a preset future time period.

10. The apparatus according to claim 6, characterized in that, The summing unit is used to sum the load components of various charging stations to obtain the total load of the multiple charging stations.

11. An electricity sales agency system, characterized in that, The electricity sales agency system includes interconnected electricity trading devices and the multi-charging station user load forecasting device as described in claim 6. The electricity sales agency system can also be connected to the electricity market system and multiple charging stations. The power trading device is used to purchase electricity from the power market system based on the total load obtained by the multi-charging station user load forecasting device, and then sell it to each charging station.

12. An electronic device, comprising: processor; as well as A memory configured to store computer-executable instructions, characterized in that, when executed, the executable instructions cause the processor to perform the multi-charging station user load forecasting method according to any one of claims 1 to 5.

13. A computer-readable storage medium storing one or more programs, characterized in that, When the one or more programs are executed by an electronic device including multiple applications, the electronic device performs the multi-charging station user load prediction method according to any one of claims 1 to 5.

Images