Charging station cluster schedulability prediction method and system
By combining the models of convolutional neural network, bidirectional long and short-term memory network and attention mechanism, a charging station cluster scheduling capability prediction model is established, which solves the problem of low prediction accuracy in the existing technology and achieves higher prediction accuracy .
Patent Information
- Application Number
- CN202411454825.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-17
- Publication Date
- 2025-02-07
AI Technical Summary
In the prior art, the accuracy of the prediction of the scheduling capability of the charging station cluster is low, mainly due to the insufficient prediction methods based on traditional statistics or a single machine learning algorithm.
A combination model of convolutional neural network, bidirectional long and short-term memory network and attention mechanism is adopted to establish a prediction model for scheduling capability of charging station clusters. This model aggregates and predicts by establishing a scheduling capability evaluation model for a single electric vehicle and combining charging station feature data.
It improves the accuracy of prediction of the scheduling capability of the charging station cluster, and can more effectively capture multi-dimensional features and the time correlation before and after, reduces the loss of key information, and improves the accuracy of the prediction results.
Smart Images
Figure CN119401405A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure relates to the technical field of power systems, and in particular to a method and system for predicting the dispatchable capacity of a charging station cluster. Background Art
[0002] The current global energy system is developing towards the direction of clean and low-carbon energy transformation, and a new power system based on new energy is gradually being established. As traditional regulation resources in the new power system are gradually replaced by intermittent distributed generation resources, electric vehicles have become an important load-side adjustable resource. Therefore, it is necessary to effectively evaluate and predict the dispatchability of charging station clusters.
[0003] In the related art, when evaluating and predicting the dispatchability of a charging station cluster, traditional prediction algorithms based on statistics or single machine learning algorithms are usually used, and the prediction accuracy is low.
[0004] Therefore, how to solve the above problems has become one of the technical problems that need to be solved urgently at this stage. Summary of the invention
[0005] In order to solve the above technical problems, the present disclosure provides a method and system for predicting the dispatchable capacity of a charging station cluster, so as to improve the accuracy of predicting the dispatchable capacity of a charging station cluster.
[0006] The present disclosure provides a method for predicting the dispatchable capacity of a charging station cluster, comprising:
[0007] Establish a dispatchability assessment model for a single electric vehicle in a charging station;
[0008] Based on the single electric vehicle dispatchability capacity evaluation model, the dispatchability capacity of the electric vehicle cluster in the charging station is aggregated to obtain the dispatchability capacity evaluation data of a single charging station;
[0009] A charging station cluster dispatchability capacity prediction model is established according to the charging station dispatchability capacity evaluation data and feature data, and the charging station cluster dispatchability capacity is predicted based on the charging station cluster dispatchability capacity prediction model. The charging station cluster dispatchability capacity prediction model adopts a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism.
[0010] Optionally, where:
[0011] The establishment of a dispatchable capacity evaluation model for a single electric vehicle in a charging station includes:
[0012] Calculating whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station, the calculation method includes:
[0013]
[0014] Where i>0, j>0, and i and j are both integers;
[0015] S i,j (t) indicates whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station;
[0016] t d i,j represents the expected departure time of the electric vehicle at the jth charging pile in the i-th charging station;
[0017] kΔt represents the initial time of the kth dispatchability evaluation period;
[0018] kΔt+Δt represents the end time of the kth dispatchability evaluation period.
[0019] Optionally, where:
[0020] The establishment of a dispatchability evaluation model for a single electric vehicle in a charging station also includes:
[0021] Calculate the maximum down-regulation power ΔP that an electric vehicle can provide EV,downi,j,k (t), the calculation method includes:
[0022]
[0023] in,
[0024] represents the maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t;
[0025] It represents the charging and discharging power of the electric vehicle at the jth charging pile in the i-th charging station under the original plan at time t.
[0026] Optionally, where:
[0027] The maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t The calculation methods include:
[0028]
[0029] in,
[0030] Indicates the maximum charging power;
[0031] Indicates the maximum discharge power;
[0032] ΔEdowni,j (kΔt) represents the difference between the minimum state of charge required to be reached by the electric vehicle at the jth charging pile in the i-th charging station at the end of the k-th dispatchable capacity assessment period and the state of charge at the initial time,
[0033]
[0034] Among them, E down i,j,k represents the minimum state of charge that the electric vehicle needs to achieve at the end of the kth dispatchability capacity evaluation period at the jth charging pile in the i-th charging station, E i,j (kΔt) represents the charge state of the electric vehicle at the jth charging pile in the i-th charging station at the initial moment of the k-th dispatchable capacity evaluation period; the calculation method of the minimum charge state required to be reached by the electric vehicle at the jth charging pile in the i-th charging station at the end moment of the k-th dispatchable capacity evaluation period includes:
[0035]
[0036] Among them, E expi,j is the expected charge state of the electric vehicle at the jth charging pile in the i-th charging station; E mini,j C is the minimum state of charge that the electric vehicle needs to reach at the jth charging pile in the i-th charging station; i,j is the battery capacity of the electric vehicle at the jth charging pile in the ith charging station.
[0037] Optionally, where:
[0038] The establishment of a dispatchability evaluation model for a single electric vehicle in a charging station also includes:
[0039] Calculating the maximum upregulated power available from an electric vehicle The calculation methods include:
[0040]
[0041] in,
[0042] It represents the maximum charging power that can actually be provided to the electric vehicle at the jth charging pile in the i-th charging station at time t.
[0043] Optionally, where:
[0044] The maximum charging power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t The calculation methods include:
[0045]
[0046] in,
[0047] Indicates the maximum charging power;
[0048] Indicates the maximum discharge power;
[0049] ΔE upi,j (kΔt) represents the charge state adjustment margin of the electric vehicle at the jth charging pile in the i-th charging station at the time kΔt,
[0050]
[0051] Among them, E expi,j is the expected state of charge of the electric vehicle at the jth charging pile in the ith charging station.
[0052] Optionally, where:
[0053] The calculation method of the dispatchable capacity evaluation data of a single charging station includes:
[0054] Calculate the maximum increase power that the i-th charging station can provide at time t and maximum power reduction The calculation methods include:
[0055]
[0056]
[0057] Calculate the maximum increase power P that all the charging stations can provide at time t CSs,up (t) and the maximum down-regulated power P CSs,down (t), the calculation method includes:
[0058]
[0059]
[0060] Calculate the upper bound of the dispatchable capacity of the i-th charging station in the period (kΔt, kΔt+Δt] (t) and the lower bound of dispatchability The calculation methods include:
[0061]
[0062]
[0063] Calculate the upper bound of the dispatchable capacity of the charging station cluster in the period (kΔt, kΔt+Δt] and the dispatchability lower bound The calculation methods include:
[0064]
[0065]
[0066] Among them, Z CSi is the total number of charging piles at the i-th charging station.
[0067] Optionally, where:
[0068] The characteristic data includes at least one of weather data, traffic data and electricity price data. The weather data includes at least one of temperature, humidity and rainfall. The traffic data includes the average value of real-time congestion of roads around the charging station. The electricity price data includes the sum of the real-time electricity price and the service fee per kilowatt-hour of the charging station.
[0069] Optionally, where:
[0070] The establishment of a charging station cluster dispatchability prediction model includes:
[0071] The charging station dispatchability evaluation data and the feature data are subjected to two convolutional layers and two pooling layers in the convolutional neural network to alternately realize feature extraction and local refinement, and the attention mechanism is included between the output end of the first convolutional layer and the output end of the second pooling layer;
[0072] Flatten the data using the flatten layer;
[0073] The processed data is input into the bidirectional long short-term memory network for training and the output data features are enhanced through the fully connected layer.
[0074] In a second aspect, based on the same inventive concept, the present disclosure further provides a charging station cluster dispatchable capacity prediction system, comprising:
[0075] The electric vehicle dispatchability assessment module is used to establish a dispatchability assessment model for a single electric vehicle in a charging station;
[0076] A charging station dispatchability assessment module is used to aggregate the dispatchability of the electric vehicle cluster in the charging station based on the single electric vehicle dispatchability assessment model to obtain dispatchability assessment data of a single charging station;
[0077] A prediction model building module, used to establish a charging station cluster dispatchability prediction model based on the charging station dispatchability evaluation data and feature data, wherein the charging station cluster dispatchability prediction model adopts a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism;
[0078] The charging station cluster dispatchable capacity prediction module is used to predict the dispatchable capacity of the charging station cluster based on the charging station cluster dispatchable capacity prediction model.
[0079] Compared with the prior art, the technical solution provided by the embodiments of the present disclosure has the following advantages:
[0080] The present disclosure provides a method and system for predicting the dispatchable capacity of a charging station cluster, wherein the prediction method includes: establishing a dispatchable capacity evaluation model for a single electric vehicle in a charging station; based on the dispatchable capacity evaluation model for a single electric vehicle, aggregating the dispatchable capacity of a cluster of electric vehicles in a charging station to obtain dispatchable capacity evaluation data for a single charging station; establishing a dispatchable capacity prediction model for a charging station cluster according to the dispatchable capacity evaluation data and feature data of the charging station, predicting the dispatchable capacity of the charging station cluster based on the dispatchable capacity prediction model of the charging station cluster, and the dispatchable capacity prediction model of the charging station cluster adopts a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism. When predicting the dispatchable capacity of a charging station cluster in a future period, the present disclosure establishes a dispatchable capacity prediction model for a charging station cluster according to the dispatchable capacity evaluation data of the charging station and in combination with the feature data, and the dispatchable capacity prediction model for a charging station cluster integrates the advantages of mining feature relationships between convolutional neural networks and mining time correlations between bidirectional long short-term memory networks, and at the same time introduces an attention mechanism to compensate for the secondary key information that may be ignored in the extraction process of the first layer of the convolutional neural network, which is conducive to effectively predicting the dispatchable capacity of a charging station cluster in a future period and improving the accuracy of the prediction. BRIEF DESCRIPTION OF THE DRAWINGS
[0081] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the present disclosure.
[0082] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.
[0083] Figure 1 The figure is a flow chart of a method for predicting the dispatchable capacity of a charging station cluster provided by an embodiment of the present disclosure;
[0084] Figure 2Shown is a schematic diagram of an actual charging plan and charging constraints provided by an embodiment of the present disclosure;
[0085] Figure 3 Shown is a schematic diagram of a charging station cluster dispatchability prediction model provided by an embodiment of the present disclosure;
[0086] Figure 4 The figure is a schematic diagram of comparison of prediction results of a maximum increased power provided by an embodiment of the present disclosure;
[0087] Figure 5 The figure is a schematic diagram of comparison of prediction results of a maximum power reduction provided by an embodiment of the present disclosure;
[0088] Figure 6 Shown is a schematic diagram of a charging station cluster dispatchable capacity prediction system provided by an embodiment of the present disclosure. DETAILED DESCRIPTION
[0089] In order to more clearly understand the above-mentioned objectives, features and advantages of the present disclosure, the scheme of the present disclosure will be further described below. It should be noted that the embodiments of the present disclosure and the features in the embodiments can be combined with each other without conflict.
[0090] In the following description, many specific details are set forth to facilitate a full understanding of the present disclosure, but the present disclosure may also be implemented in other ways different from those described herein; it is obvious that the embodiments in the specification are only part of the embodiments of the present disclosure, rather than all of the embodiments.
[0091] The inventor found in the research that electric vehicles, as a means of transportation, are special compared to general energy storage devices. They are affected by users' travel and charging needs and have strong randomness, which needs to be fully considered when evaluating and calculating the dispatchable capacity. In addition, the charging behavior of electric vehicles is affected by various external factors. The dispatchable capacity of charging station clusters is explored around a single dimension, lacking overall consideration of multiple external factors. At the same time, the existing technology for predicting the dispatchable capacity of charging station clusters is usually based on traditional statistical prediction algorithms or single machine learning algorithms. Based on the above reasons, the existing prediction methods for predicting the dispatchable capacity of charging station clusters have low prediction accuracy.
[0092] Therefore, how to solve the above problems has become one of the technical problems that need to be solved urgently at this stage.
[0093] In order to solve the above technical problems, the present disclosure provides a method and system for predicting the dispatchable capacity of a charging station cluster, so as to improve the accuracy of predicting the dispatchable capacity of a charging station cluster.
[0094] Figure 1FIG. 1 is a flow chart of a method for predicting the dispatchable capacity of a charging station cluster provided by an embodiment of the present disclosure. Please refer to Figure 1 The present disclosure provides a method for predicting the dispatchable capacity of a charging station cluster, comprising:
[0095] Step S01, establishing a dispatchability assessment model for a single electric vehicle in a charging station;
[0096] Step S02: Based on the single electric vehicle dispatchability assessment model, the dispatchability of the electric vehicle cluster in the charging station is aggregated to obtain the dispatchability assessment data of a single charging station;
[0097] Step S03: establishing a charging station cluster dispatchability prediction model according to the charging station dispatchability evaluation data and feature data, and predicting the charging station cluster dispatchability based on the charging station cluster dispatchability prediction model, wherein the charging station cluster dispatchability prediction model adopts a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism.
[0098] It should be noted that the present disclosure provides a method for predicting the dispatchable capacity of a charging station cluster, including but not limited to steps S01 to S03.
[0099] Specifically, when predicting the dispatchable capacity of a charging station cluster, it is necessary to predict the dispatchable capacity of an electric vehicle cluster in a single charging station; when predicting the dispatchable capacity of an electric vehicle cluster in a single charging station, it is necessary to predict the dispatchable capacity of a single electric vehicle in the charging station. Therefore, in step S01, the dispatchable capacity of a single electric vehicle in the charging station is first predicted, and a dispatchable capacity evaluation model for a single electric vehicle is established.
[0100] In step S02, based on the single electric vehicle dispatchable capacity assessment model established in step S01, the dispatchable capacity of the electric vehicle cluster in the charging station is aggregated to obtain the dispatchable capacity assessment data of a single charging station. It should be noted that in this step, the dispatchable assessment capacity data of the charging station is obtained, which is subsequently used as the data basis for predicting the dispatchable capacity of the charging station cluster. It should be noted that the present disclosure is only explained by taking the dispatchable capacity assessment data of a single charging station as an example, and is not limited to this. It is also possible to obtain the dispatchable capacity assessment data of each charging station for multiple different charging stations to increase the data volume, thereby improving the accuracy of the prediction.
[0101] In step S03, according to the charging station dispatchable capacity evaluation data obtained in step S02 and adding feature data, a charging station cluster dispatchable capacity prediction model is established, and the dispatchable capacity of the charging station cluster is predicted based on the charging station cluster dispatchable capacity prediction model. The charging station cluster dispatchable capacity prediction model provided by the present invention integrates the advantages of convolutional neural networks and bidirectional long short-term memory networks, and introduces an attention mechanism to strengthen the mining of multi-dimensional features and time correlations, while reducing the loss of key information in the input features, which is conducive to improving the accuracy of the prediction.
[0102] It should be noted that the Convolutional Neural Network (CNN) is a deep learning model designed specifically for processing data with a grid structure. The Convolutional Neural Network includes a convolutional layer, a pooling layer, and a fully connected layer. The convolutional layer includes multiple convolution kernels, which are small weight matrices. The pooling layer mainly performs dimensionality reduction on the data output by the convolutional layer, reduces the amount of data calculation, and improves the computational efficiency of the model. The neurons in the fully connected layer are connected to all neurons in the previous layer. The fully connected layer is used to integrate and classify the features extracted by other layers. Bidirectional Long-Short Term Memory (BiLSTM) is an improvement on Long-Short Term Memory (LSTM). It has unique advantages in processing sequence data. Bidirectional LSTM includes forward LSTM and backward LSTM. The forward LSTM processes the input sequence in the normal order of the sequence, and the backward LSTM processes the input sequence in the reverse order. The bidirectional structure of Bidirectional LSTM can fully capture the effective information in the data, which is conducive to improving the accuracy of the task. Attention Mechanism is an important technology in deep learning. Its main function is to allow the model to selectively focus on certain parts of the input data, thereby improving the performance and effect of the model.
[0103] It can be understood that the present disclosure establishes a charging station cluster dispatchability prediction model based on the charging station dispatchability evaluation data and the characteristic data. The charging station cluster dispatchability prediction model integrates the advantages of the convolutional neural network to mine the characteristic relationship and the bidirectional long short-term memory network to mine the time correlation before and after. At the same time, the attention mechanism is introduced to make up for the secondary key information that may be ignored in the extraction process of the first layer of the convolutional neural network, which is conducive to the effective prediction of the dispatchability of the charging station cluster in the future period and improves the accuracy of the prediction, thereby providing a reference for the formulation of the charging plan for the electric vehicle cluster in the charging station according to the prediction results. In addition, the charging station cluster dispatchability prediction method provided by the present disclosure is conducive to timely updating the prediction results of the charging station cluster dispatchability, which is conducive to the dispatch center issuing dispatch instructions within the response period to adjust the charging plan for the electric vehicle cluster in the charging station.
[0104] Please continue to refer to Figure 1 The present disclosure provides an optional implementation method for establishing a dispatchability evaluation model for a single electric vehicle in a charging station, including:
[0105] Calculate whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station. The calculation method includes:
[0106]
[0107] Where i>0, j>0, and i and j are both integers;
[0108] S i,j (t) indicates whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station;
[0109] t d i,j represents the expected departure time of the electric vehicle at the jth charging pile in the i-th charging station;
[0110] kΔt represents the initial time of the kth dispatchability evaluation period;
[0111] kΔt+Δt represents the end time of the kth dispatchability evaluation period.
[0112] Specifically, assuming that an electric vehicle is connected to a charging pile and connected to the grid, a parameter is needed to reflect that there is a single electric vehicle dispatchable resource that can be aggregated. Therefore, in this implementation, a dispatchable evaluation model for a single electric vehicle in a charging station is established. It is necessary to calculate whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station. S i,j (t) is a parameter that reflects whether there are EV dispatchable resources that can be aggregated at the jth charging pile in the i-th charging station. From formula (1), it can be obtained that when there is an EV connected to the grid at the jth charging pile in the i-th charging station, Si,j (t) = 1; when there is no electric vehicle connected to the grid at the jth charging pile in the i-th charging station, S i,j (t)=0.
[0113] Figure 2 FIG. 1 is a schematic diagram of an actual charging plan and charging constraints provided by an embodiment of the present disclosure. Please refer to FIG. Figure 1 and Figure 2 The present disclosure provides an optional implementation method of establishing a dispatchability evaluation model for a single electric vehicle in a charging station, further comprising:
[0114] Calculate the maximum down-regulation power ΔP that an electric vehicle can provide EV,downi,j,k (t), the calculation method includes:
[0115]
[0116] in,
[0117] It represents the maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t;
[0118] represents the charging and discharging power of the electric vehicle at the jth charging pile in the i-th charging station at time t as originally planned;
[0119] S i,j (t) indicates whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station.
[0120] Specifically, the premise for predicting the dispatchable capacity of the charging station cluster is to evaluate the controllable boundaries of the charging and discharging power of electric vehicles, including the maximum down-regulation power and maximum up-regulation power that electric vehicles can actually provide. The up-regulation power is the charging power that the electric vehicle cluster can increase, and the frequency down-regulation power is the charging power that the electric vehicle cluster can reduce, which is limited by the equipment, user demand and battery capacity. The method for predicting the dispatchable capacity of the charging station cluster provided by the present disclosure takes into account the limitations of equipment, user demand and battery capacity, which will be explained in detail in the subsequent embodiments.
[0121] It should be noted that the establishment of the evaluation model for the dispatchable capacity of electric vehicles in the charging station is to evaluate the dispatchable capacity of each electric vehicle in the charging station within the time (kΔt, kΔt+Δt] every Δt, where k represents the kth frequency regulation capacity evaluation; represents the length of the evaluation period; optionally, Δt is taken as 15 minutes. This application is only used as an example for illustration and is not limited to this.
[0122] In this embodiment, a maximum down-regulation power ΔP that an electric vehicle can provide is provided. EV,downi,j,kThe calculation method of (t) is based on the maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t. The charging and discharging power of the electric vehicle at the jth charging pile in the i-th charging station at the originally planned time t The maximum power reduction ΔP that an electric vehicle can provide can be calculated EV,downi,j,k (t). It can be concluded from formula (2) that when the jth charging pile in the i-th charging station is at the time t of the k-th dispatchability evaluation and there is no electric vehicle connected to the grid, S i,j (t) = 0, therefore, ΔP EV,downi,j,k (t)=0.
[0123] Please continue to refer to Figure 1 and Figure 2 , for the maximum discharge power that an electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t The present disclosure provides an optional implementation method in which the maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t is The calculation methods include:
[0124]
[0125] in,
[0126] Indicates the maximum charging power;
[0127] Indicates the maximum discharge power;
[0128] ΔE downi,j (kΔt) represents the difference between the minimum state of charge required to be achieved by the electric vehicle at the jth charging pile in the i-th charging station at the end of the k-th dispatchability capacity assessment period and the state of charge at the initial time.
[0129]
[0130] Among them, E down i,j,k It represents the minimum state of charge that the electric vehicle needs to reach at the end of the kth dispatchability capacity evaluation period at the jth charging pile in the i-th charging station, E i,j (kΔt) represents the charge state of the electric vehicle at the jth charging pile in the i-th charging station at the initial moment of the k-th dispatchable capacity evaluation period; the calculation method of the minimum charge state required to be reached by the electric vehicle at the jth charging pile in the i-th charging station at the end moment of the k-th dispatchable capacity evaluation period includes:
[0131]
[0132] Among them, E expi,j is the expected state of charge of the electric vehicle at the jth charging pile in the i-th charging station; E mini,j is the minimum state of charge that the electric vehicle needs to reach at the jth charging pile in the i-th charging station; C i,j is the battery capacity of the electric vehicle at the jth charging pile in the i-th charging station.
[0133] Specifically, in order to ensure that the electric vehicle can reach the user's expected state of charge when leaving the station, the minimum state of charge that the electric vehicle needs to reach at the end time kΔt+Δt of the kth dispatchable capacity evaluation period is defined as E down i,j,k In this regard, electric vehicles are divided into two types: those that remain connected to the grid and those that are about to go off the grid during the kth dispatchability assessment period. Then, the minimum state of charge E that the electric vehicle at the jth charging pile in the i-th charging station needs to reach at the end of the kth dispatchability assessment period can be calculated using formula (5): down i,j,k If the electric vehicle at the jth charging pile in the i-th charging station remains connected to the grid during the k-th dispatchability evaluation period, at the final time kΔt+Δt, if the state of charge is not less than E min i,j , and then at the departure time t d i,j If it continues to charge at the maximum charging power, it can reach E expi,j If you leave the station, you need to leave at the departure time t d i,j Before reaching E expi,j .
[0134] The difference ΔE between the minimum state of charge required to be reached by the electric vehicle at the jth charging pile in the i-th charging station at the end of the k-th dispatchable capacity assessment period and the state of charge at the initial time is calculated by formula (4): downi,j (kΔt) is the charge margin of the electric vehicle at the jth charging pile in the i-th charging station at the time kΔt. Furthermore, if the electric vehicle at the jth charging pile in the i-th charging station remains connected to the grid during the k-th dispatchability evaluation period, its grid connection time during the evaluation period is Δt; if it leaves the station during the period, its grid connection time is t d i,j -kΔt. At the same time, considering the maximum charging power due to equipment limitations and maximum discharge power The maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t is calculated by formula (3):
[0135] It should be noted that the embodiment of the present disclosure fully considers the user's charging needs and equipment limitations when establishing the dispatchability assessment model of a single electric vehicle in a charging station, which is further conducive to improving the accuracy of the prediction of the dispatchability of the charging station cluster.
[0136] Please continue to refer to Figure 1 and Figure 2 The present disclosure provides an optional implementation method of establishing a dispatchability evaluation model for a single electric vehicle in a charging station, further comprising:
[0137] Calculating the maximum upregulated power available from an electric vehicle The calculation methods include:
[0138]
[0139] in,
[0140] It represents the maximum charging power that can actually be provided to the electric vehicle at the jth charging pile in the i-th charging station at time t.
[0141] Specifically, in this embodiment, a maximum power increase that an electric vehicle can provide is provided. The calculation method is based on the maximum charging power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t. The charging and discharging power of the electric vehicle at the jth charging pile in the i-th charging station at the originally planned time t The maximum upregulated power that an electric vehicle can provide can be calculated From formula (6), it can be concluded that when the jth charging pile in the i-th charging station is at time t and there is no electric vehicle connected to the grid, ΔP EV,up i,j (t)=0.
[0142] Please continue to refer to Figure 1 and Figure 2 , the maximum charging power that can be provided by the electric vehicle at the jth charging pile in the i-th charging station at time t The present disclosure provides an optional implementation method in which the maximum charging power that can actually be provided by the electric vehicle at the jth charging pile in the i-th charging station at time t is The calculation methods include:
[0143]
[0144] in,
[0145] Indicates the maximum charging power;
[0146] Indicates the maximum discharge power;
[0147] ΔE upi,j (kΔt) represents the charge state adjustment margin of the electric vehicle at the jth charging pile in the i-th charging station at time kΔt,
[0148]
[0149] Among them, E expi,j is the expected state of charge of the electric vehicle at the jth charging pile in the i-th charging station.
[0150] Specifically, considering the battery capacity limit of electric vehicles, the charge state adjustment margin of the electric vehicle at the jth charging pile in the i-th charging station at time kΔt is set to ΔE upi,j (kΔt), the upward adjustment margin ΔE can be calculated by formula (8): upi,j (kΔt). Further, considering the maximum charging power due to device limitations and maximum discharge power As well as the two cases of electric vehicles connected to the grid and off the grid, the maximum charging power that can be actually provided by the electric vehicle at the jth charging pile in the i-th charging station at time t can be calculated by formula (7):
[0151] It should be noted that, when establishing the dispatchability assessment model for a single electric vehicle in a charging station, the disclosed embodiment also takes into account the battery capacity limitation of the electric vehicle, which is further beneficial to improving the accuracy of the prediction of the dispatchability of the charging station cluster.
[0152] Please refer to Figure 1 The present disclosure provides an optional implementation mode in which the method for calculating the dispatchable capacity evaluation data of a single charging station in step S02 includes:
[0153] Calculate the maximum increase in power that the i-th charging station can provide at time t and maximum power reduction The calculation methods include:
[0154]
[0155]
[0156] Calculate the maximum increase power P that all charging stations can provide at time t CSs,up (t) and the maximum down-regulated power P CSs ,down (t), the calculation method includes:
[0157]
[0158]
[0159] Calculate the upper bound of the dispatchable capacity of the i-th charging station in the period (kΔt, kΔt+Δt] and the dispatchability lower bound The calculation methods include:
[0160]
[0161]
[0162] Calculate the upper bound of the dispatchable capacity of the charging station cluster in the period (kΔt, kΔt+Δt] and the dispatchability lower bound The calculation methods include:
[0163]
[0164]
[0165] Among them, Z CSi is the total number of charging piles at the i-th charging station.
[0166] Specifically, in the method for predicting the dispatchable capacity of a charging station cluster provided in the embodiment of the present disclosure, a dispatchable capacity evaluation model for a single electric vehicle in a charging station is established in step S01. In step S02, based on the evaluation of the dispatchable capacity of a single electric vehicle, the dispatchable capacity of the electric vehicle cluster in the charging station can be aggregated, and the maximum up-regulation power that can be provided by the i-th charging station at time t is calculated by formula (9) to formula (16). and maximum power reduction The maximum increase power P that all charging stations can provide at time t CSs,up (t) and the maximum down-regulated power P CSs ,down (t), the upper bound of the dispatchable capacity of the i-th charging station in the period (kΔt, kΔt+Δt] and the dispatchability lower bound And the upper bound of the dispatchable capacity of the charging station cluster in the period (kΔt, kΔt+Δt] and the dispatchability lower bound (t), the above data obtained will be used as the data basis for predicting the dispatchable capacity of the charging station cluster.
[0167] Please continue to refer to Figure 1The present disclosure provides an optional implementation mode in which the characteristic data includes at least one of weather data, traffic data and electricity price data. The weather data includes at least one of temperature, humidity and rainfall. The traffic data includes the average value of the real-time congestion of roads around the charging station. The electricity price data includes the sum of the real-time electricity price and the service fee per kilowatt-hour of the charging station.
[0168] Specifically, the method for predicting the dispatchable capacity of a charging station cluster provided by the present disclosure also sets characteristic data when establishing a prediction model for the dispatchable capacity of a charging station cluster, and the characteristic data includes at least one of weather data, traffic data, and electricity price data. Exemplarily, a certain charging station is selected as a research object, and when establishing a prediction model for the dispatchable capacity of a charging station cluster, the collected characteristic data such as weather data, traffic data, and electricity price data are input into the prediction model for the dispatchable capacity of the charging station cluster together with the dispatchable capacity evaluation data of the charging station. Such a setting is more conducive to improving the accuracy of the prediction of the dispatchable capacity of the charging station cluster.
[0169] It should be noted that the time period for obtaining weather data, traffic data and electricity price data can be 15 minutes, but this application only uses this as an example for illustration and is not limited to this. For example, it can also be 20 minutes, 30 minutes, etc.
[0170] Figure 3 FIG. 1 is a schematic diagram of a charging station cluster dispatchability prediction model provided by an embodiment of the present disclosure. Please refer to Figure 1 and Figure 3 The present disclosure provides an optional implementation mode, wherein establishing a charging station cluster dispatchability prediction model includes:
[0171] The charging station dispatchability assessment data and feature data are passed through two convolutional layers and two pooling layers in the convolutional neural network to alternately implement feature extraction and local refinement. An attention mechanism is included between the output of the first convolutional layer and the output of the second pooling layer.
[0172] Flatten the data using the flatten layer;
[0173] The processed data is input into the bidirectional long short-term memory network for training and the output data features are enhanced through the fully connected layer.
[0174] Specifically, this embodiment provides a method for establishing a charging station cluster dispatchable capacity prediction model. First, the charging station dispatchable evaluation data calculated in step S02 and feature data including weather data, traffic data and electricity price data are input. The above data is subjected to two layers of continuously deepening one-dimensional convolutional layers and pooling layers in the convolutional neural network to alternately realize feature extraction and local refinement. Optionally, the number of convolution kernels is set to 32, 64, and the step size is set to 3; an attention mechanism is also set between the output end of the first convolutional layer and the output end of the second pooling layer, which is conducive to avoiding the loss of key information. The data is then flattened through the flattening layer; the processed data is input into a bidirectional long short-term memory network for training, and optionally, the number of neurons is 128. Finally, the training results or prediction results of the dispatchable capacity of the charging station cluster are passed through the fully connected layer to enhance the output data features.
[0175] In order to verify the accuracy of the charging station cluster dispatchability prediction model provided by the present disclosure, the present disclosure also compares the charging station cluster dispatchability prediction model with the LSTM model, the BiLST model and the CNN-BiLSTM model, wherein the LSTM model is a model formed by only using the long short-term memory network, the BiLST model is a model formed by only using the bidirectional long short-term memory network, and the CNN-BiLSTM model is a model formed by using the convolutional neural network and the bidirectional long short-term memory network. The charging station cluster dispatchability prediction model provided by the present disclosure is illustrated as CNN-LSTM-Attention in the figure. For the prediction results of the maximum increase and maximum decrease power that can be provided by a charging station cluster consisting of 10 charging stations in a certain area, please refer to Figure 4 and Figure 5 , Figure 4 FIG. 1 is a schematic diagram showing a comparison of prediction results of a maximum increased power provided by an embodiment of the present disclosure. Figure 5 The figure shows a comparison diagram of the prediction results of the maximum power reduction provided by an embodiment of the present disclosure. It can be seen from the figure that the prediction accuracy of the charging station cluster dispatchable capacity prediction model provided by the present disclosure is higher.
[0176] It should be noted that the batch size and training epochs of the built CNN-BiLSTM model are set to 32 and 60 respectively, and the mean square error MSE is selected in conjunction with the Adam optimizer to adjust the model parameters. In addition, this paper uses the mean square error MSE and the average percentage error MAPE as evaluation indicators for the prediction of the adjustable frequency capacity of the charging station. The calculation methods are shown in formulas (17) and (18) respectively:
[0177]
[0178]
[0179] Among them, y i,m is the actual value of the dispatchable capacity of the i-th charging station at the m-th sample point; y pi,m is the predicted value of the dispatchable capacity of the charging station cluster of the i-th charging station in the m-th sample point; N is the total number of sample points.
[0180] Based on the same inventive concept, the present disclosure also provides a charging station cluster dispatchable capacity prediction system. Figure 6 FIG. 1 is a schematic diagram of a charging station cluster dispatchable capacity prediction system provided by an embodiment of the present disclosure. The charging station cluster dispatchable capacity prediction system 100 includes:
[0181] The electric vehicle dispatchability evaluation module 10 is used to establish a dispatchability evaluation model for a single electric vehicle in a charging station;
[0182] The charging station dispatchability evaluation module 20 is used to aggregate the dispatchability of the electric vehicle cluster in the charging station based on the single electric vehicle dispatchability evaluation model to obtain the dispatchability evaluation data of the single charging station;
[0183] A prediction model building module 30 is used to build a charging station cluster dispatchability prediction model based on the charging station dispatchability evaluation data and feature data. The charging station cluster dispatchability prediction model adopts a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism.
[0184] The charging station cluster dispatchable capacity prediction module 40 is used to predict the dispatchable capacity of the charging station cluster based on the charging station cluster dispatchable capacity prediction model.
[0185] It should be noted that the charging station cluster dispatchable capacity prediction system 100 provided by the present disclosure adopts the charging station cluster dispatchable capacity prediction method provided by the present disclosure to predict the dispatchable capacity of the charging station cluster. The specific embodiment of the charging station cluster dispatchable capacity prediction method can refer to the above content. The present disclosure establishes a charging station cluster dispatchable capacity prediction model based on the charging station dispatchable capacity evaluation data and the feature data. The charging station cluster dispatchable capacity prediction model integrates the advantages of the convolutional neural network mining feature relationship and the bidirectional long short-term memory network mining before and after time correlation. At the same time, the attention mechanism is introduced to make up for the secondary key information that may be ignored in the extraction process of the first layer of the convolutional neural network, which is conducive to the effective prediction of the dispatchable capacity of the charging station cluster in the future period, and improves the accuracy of the prediction, so as to provide a reference for the formulation of the charging plan of the electric vehicle cluster in the charging station according to the prediction results. In addition, the charging station cluster dispatchable capacity prediction method provided by the present disclosure is conducive to timely updating the prediction results of the dispatchable capacity of the charging station cluster, which is conducive to the dispatch center issuing dispatch instructions within the response period to achieve the adjustment of the charging plan of the electric vehicle cluster in the charging station.
[0186] It can be seen from the above embodiments that the method and system for predicting the dispatchable capacity of a charging station cluster provided by the present disclosure achieve at least the following beneficial effects:
[0187] The present disclosure provides a method and system for predicting the dispatchable capacity of a charging station cluster, wherein the prediction method includes: establishing a dispatchable capacity evaluation model for a single electric vehicle in a charging station; based on the dispatchable capacity evaluation model for a single electric vehicle, aggregating the dispatchable capacity of a cluster of electric vehicles in a charging station to obtain dispatchable capacity evaluation data for a single charging station; establishing a dispatchable capacity prediction model for a charging station cluster according to the dispatchable capacity evaluation data and feature data of the charging station, predicting the dispatchable capacity of the charging station cluster based on the dispatchable capacity prediction model of the charging station cluster, and the dispatchable capacity prediction model of the charging station cluster adopts a convolutional neural network, a bidirectional long short-term memory network, and an attention mechanism. When predicting the dispatchable capacity of a charging station cluster in a future period, the present disclosure establishes a dispatchable capacity prediction model for a charging station cluster according to the dispatchable capacity evaluation data of the charging station and in combination with the feature data, and the dispatchable capacity prediction model for a charging station cluster integrates the advantages of mining feature relationships between convolutional neural networks and mining time correlations between bidirectional long short-term memory networks, and at the same time introduces an attention mechanism to compensate for the secondary key information that may be ignored in the extraction process of the first layer of the convolutional neural network, which is conducive to effectively predicting the dispatchable capacity of a charging station cluster in a future period and improving the accuracy of the prediction.
[0188] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0189] The above description is only a specific embodiment of the present disclosure, so that those skilled in the art can understand or implement the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present disclosure. Therefore, the present disclosure will not be limited to the embodiments described herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for predicting the dispatchable capacity of a charging station cluster, characterized in that: include: Establish a dispatchability assessment model for a single electric vehicle in a charging station; Based on the single electric vehicle dispatchability capacity evaluation model, the dispatchability capacity of the electric vehicle cluster in the charging station is aggregated to obtain the dispatchability capacity evaluation data of a single charging station; A charging station cluster dispatchability capacity prediction model is established according to the charging station dispatchability capacity evaluation data and feature data, and the charging station cluster dispatchability capacity is predicted based on the charging station cluster dispatchability capacity prediction model. The charging station cluster dispatchability capacity prediction model adopts a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism.
2. The method for predicting the dispatchable capacity of a charging station cluster according to claim 1, characterized in that: The establishment of a dispatchable capacity evaluation model for a single electric vehicle in a charging station includes: Calculating whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station, the calculation method includes: Where i>0, j>0, and i and j are both integers; S i,j (t) indicates whether there is an electric vehicle connected to the grid at the jth charging pile in the i-th charging station; t di,j represents the expected departure time of the electric vehicle at the jth charging pile in the i-th charging station; kΔt represents the initial time of the kth dispatchability evaluation period; kΔt+Δt represents the end time of the kth dispatchability evaluation period.
3. The method for predicting the dispatchable capacity of a charging station cluster according to claim 2, characterized in that: The establishment of a dispatchability evaluation model for a single electric vehicle in a charging station also includes: Calculate the maximum down-regulation power ΔP that an electric vehicle can provide EV,downi,j,k (t), the calculation method includes: in, represents the maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t; It represents the charging and discharging power of the electric vehicle at the jth charging pile in the i-th charging station under the original plan at time t.
4. The method for predicting the dispatchable capacity of a charging station cluster according to claim 3 is characterized in that: The maximum discharge power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t The calculation methods include: in, Indicates the maximum charging power; Indicates the maximum discharge power; ΔE downi,j (kΔt) represents the difference between the minimum state of charge required to be reached by the electric vehicle at the jth charging pile in the i-th charging station at the end of the k-th dispatchable capacity assessment period and the state of charge at the initial time, Among them, E downi,j,k represents the minimum state of charge that the electric vehicle needs to achieve at the end of the kth dispatchability capacity evaluation period at the jth charging pile in the i-th charging station, E i,j (kΔt) represents the charge state of the electric vehicle at the jth charging pile in the i-th charging station at the initial moment of the k-th dispatchable capacity evaluation period; the calculation method of the minimum charge state required to be reached by the electric vehicle at the jth charging pile in the i-th charging station at the end moment of the k-th dispatchable capacity evaluation period includes: Among them, E expi,j is the expected charge state of the electric vehicle at the jth charging pile in the i-th charging station; E mini,j C is the minimum state of charge that the electric vehicle needs to reach at the jth charging pile in the i-th charging station; i,j is the battery capacity of the electric vehicle at the jth charging pile in the ith charging station.
5. The method for predicting the dispatchable capacity of a charging station cluster according to claim 3, characterized in that: The establishment of a dispatchability evaluation model for a single electric vehicle in a charging station also includes: Calculating the maximum upregulated power available from an electric vehicle The calculation methods include: in, It represents the maximum charging power that can actually be provided to the electric vehicle at the jth charging pile in the i-th charging station at time t.
6. The method for predicting the dispatchable capacity of a charging station cluster according to claim 5, characterized in that: The maximum charging power that the electric vehicle can actually provide at the jth charging pile in the i-th charging station at time t The calculation methods include: in, Indicates the maximum charging power; Indicates the maximum discharge power; ΔE upi,j (kΔt) represents the charge state adjustment margin of the electric vehicle at the jth charging pile in the i-th charging station at the time kΔt, Among them, E expi,j is the expected state of charge of the electric vehicle at the jth charging pile in the i-th charging station.
7. The method for predicting the dispatchable capacity of a charging station cluster according to claim 5, characterized in that: The calculation method of the dispatchable capacity evaluation data of a single charging station includes: Calculate the maximum increase power that the i-th charging station can provide at time t and maximum power reduction The calculation methods include: Calculate the maximum increase power P that all the charging stations can provide at time t CSs,up (t) and the maximum down-regulated power P CSs ,down (t), the calculation method includes: Calculate the upper bound of the dispatchable capacity of the i-th charging station in the period (kΔt, kΔt+Δt] (t) and the lower bound of dispatchability The calculation methods include: Calculate the upper bound of the dispatchable capacity of the charging station cluster in the period (kΔt, kΔt+Δt] and the dispatchability lower bound The calculation methods include: Among them, Z CSi is the total number of charging piles at the i-th charging station.
8. The method for predicting the dispatchable capacity of a charging station cluster according to claim 1, characterized in that: The characteristic data includes at least one of weather data, traffic data and electricity price data. The weather data includes at least one of temperature, humidity and rainfall. The traffic data includes the average value of real-time congestion of roads around the charging station. The electricity price data includes the sum of the real-time electricity price and the service fee per kilowatt-hour of the charging station.
9. The method for predicting the dispatchable capacity of a charging station cluster according to claim 8, characterized in that: The establishment of a charging station cluster dispatchability prediction model includes: The charging station dispatchability evaluation data and the feature data are subjected to two convolutional layers and two pooling layers in the convolutional neural network to alternately realize feature extraction and local refinement, and the attention mechanism is included between the output end of the first convolutional layer and the output end of the second pooling layer; Flatten the data using the flatten layer; The processed data is input into the bidirectional long short-term memory network for training and the output data features are enhanced through the fully connected layer.
10. A charging station cluster dispatchability prediction system, characterized in that: include: The electric vehicle dispatchability assessment module is used to establish a dispatchability assessment model for a single electric vehicle in a charging station; A charging station dispatchability assessment module is used to aggregate the dispatchability of the electric vehicle cluster in the charging station based on the single electric vehicle dispatchability assessment model to obtain dispatchability assessment data of a single charging station; A prediction model building module, used to establish a charging station cluster dispatchability prediction model based on the charging station dispatchability evaluation data and feature data, wherein the charging station cluster dispatchability prediction model adopts a convolutional neural network, a bidirectional long short-term memory network and an attention mechanism; The charging station cluster dispatchable capacity prediction module is used to predict the dispatchable capacity of the charging station cluster based on the charging station cluster dispatchable capacity prediction model.