Electric vehicle frequency modulation participation service system based on cloud edge collaborative architecture

Abstract

The invention provides an electric vehicle frequency modulation participation service system based on a cloud-side collaborative architecture. The system comprises terminal equipment, edge equipment, a cloud end and an automatic power generation control master station system. According to the electric vehicle participation frequency modulation service system based on the cloud-side collaborative architecture, the frequency modulation capacity of each charging station in the next response time period is predicted in a rolling mode in real time based on the cloud-side collaborative regulation and control architecture, the frequency modulation capacity is used as the basis of instruction decomposition between the charging stations, the risk of time delay between different levels of equipment is reduced, and the service efficiency is improved. The charging demand of the electric vehicle user on the time scale except the expected electric quantity is preferentially considered, and it is ensured that the electric vehicle participates in frequency modulation within the time range expected by the user. Through layer-by-layer cooperation of the cloud, the edge device and the terminal device, the electric vehicle can participate in the frequency modulation service, and meanwhile, the requirements of a power grid, an aggregator and a user are met as much as possible.

Description

Technical Field

[0001] The present application belongs to the field of power grid technology, and more specifically, to a frequency regulation service system for electric vehicles based on a cloud-edge collaborative architecture. Background Art

[0002] Electric vehicles are considered as typical flexible loads in new power systems. Compared with traditional power generation devices, the advantages of electric vehicles in frequency regulation services are obvious, including rapid response, strong controllability, and high cost-effectiveness. However, due to the small power of a single electric vehicle and random grid connection, it is quite challenging to directly control a single electric vehicle to participate in frequency regulation.

[0003] To solve this problem, current research and practice generally use electric vehicle aggregators as intermediaries to undertake the tasks of load integration and dispatch control. In order to improve the reliability and effectiveness of the frequency regulation services provided by electric vehicle aggregators, how to accurately evaluate the frequency regulation capacity of electric vehicles and quickly and optimally decompose frequency regulation instructions has become a hot topic for electric vehicles to participate in frequency regulation services.

[0004] Current methods for evaluating frequency regulation capacity include building a probability distribution model by fitting historical data and using Monte Carlo simulation to evaluate the frequency regulation capacity of electric vehicles. Some methods also use random variables to describe the uncertainty of the frequency regulation capacity of electric vehicles and obtain the probability distribution of the frequency regulation capacity of electric vehicles based on non-parametric estimation methods. Some scholars have also proposed an electric vehicle frequency regulation capacity evaluation method based on a cloud-edge fusion architecture, with real-time rolling evaluation and layer-by-layer aggregation.

[0005] In the decomposition of frequency modulation instructions, the existing foreign frequency modulation performance compensation mechanism is used as the background, and a robust optimization method is used to ensure that electric vehicles respond to random frequency modulation instructions in a timely manner, while ensuring the maximization of profits on the electric vehicle aggregator side; it also considers the charging needs of electric vehicle users, tracks the expected charging status of electric vehicles when participating in frequency modulation services in real time, and formulates an optimization strategy for instruction allocation.

[0006] The prior art has the following disadvantages:

[0007] (1) In terms of frequency modulation capacity evaluation methods, the probability distribution-based evaluation method lacks the quantitative calculation of the frequency modulation capacity of electric vehicles in terms of time series, and the evaluation results are less accurate. Although the existing evaluation method using the cloud-edge collaborative architecture improves the evaluation accuracy, it ignores the latency between devices at different levels under the cloud-edge fusion framework, resulting in the decomposition of instructions being completely dependent on real-time evaluation results, which is not conducive to the rapid decomposition of frequency modulation instructions.

[0008] (2) In terms of the decomposition of frequency modulation instructions, the above-mentioned optimization decomposition methods mostly rely on the policies of foreign power markets, and rarely consider the establishment of models applicable to the Chinese power market, making it difficult to apply them to actual projects. In addition, they do not consider how to better improve the comprehensive satisfaction of the three parties: the power grid, electric vehicle aggregators, and electric vehicle users. In particular, in terms of electric vehicle users, existing research only considers meeting the user's target power consumption, and does not consider the user's charging needs on a time scale. Summary of the invention

[0009] In order to solve the above technical problems, the technical solution adopted in this application is to provide an electric vehicle frequency modulation service system based on a cloud-edge collaborative architecture, including:

[0010] Terminal device: used to set the user's charging requirements and upload the real-time collected charging pile status information to the edge device;

[0011] Edge devices: used to determine whether electric vehicles can meet the charging requirements set by users during the period of frequency modulation; rolling calculation of the frequency modulation capacity of electric vehicles in the charging station based on the status information of the charging piles and uploading the calculation results to the cloud; receiving the frequency modulation instructions issued by the cloud to the charging station and decomposing them, including: comprehensively adjusting the frequency modulation instructions of each charging pile in the station according to the frequency modulation instructions issued by the cloud, and issuing them to each charging pile;

[0012] Cloud: used to predict the frequency regulation load of the subordinate charging stations according to the calculation results of the edge devices, and upload it to the automatic power generation control master station system; receive the general frequency regulation instruction issued by the automatic power generation control master station system, and decompose it, including: adjusting the frequency regulation instructions of the subordinate charging stations according to the general frequency regulation instruction issued by the automatic power generation control master station system, and issuing it to each charging station;

[0013] Automatic power generation control master station system: sends general frequency regulation instructions to the cloud based on the forecast results of frequency regulation load.

[0014] Optionally, the method for determining whether the charging demand set by the user meets the frequency modulation condition is as follows:

[0015]

[0016] Where: y c =1 and y c =0 means that the charging demand of the user can be met or cannot be met respectively; The maximum power of an electric vehicle charging at the kth charging pile in the charging station; The maximum charging time set for users charging at the kth charging pile in the charging station; a is the minimum threshold coefficient for judging whether the demand can be met; The expected amount of charge to be charged by the user charging at the kth charging pile in the charging station is calculated as follows:

[0017]

[0018] Where: The battery capacity of the electric vehicle charged at the kth charging pile in the charging station; The expected power set for the user charging at the kth charging pile in the charging station; is the power of the electric vehicle charged at the kth charging pile in the charging station at the time it0; i represents the ith frequency modulation period; it0 is the starting time of the ith response period; where i is a positive integer greater than or equal to 1 and less than or equal to 96, and t0 is the frequency modulation capacity evaluation time interval;

[0019] If the judgment result is y c = 0, the user will reset the maximum allowed charging time Expected power Until y c =1.

[0020] Optionally, the method for calculating the frequency regulation capacity of electric vehicles in a charging station is as follows:

[0021] Calculate the increase in frequency regulation capacity of a single electric vehicle:

[0022]

[0023] Where: The upper limit of the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period;

[0024] Calculate the frequency regulation capacity increase capacity based on the frequency regulation capacity upper limit:

[0025]

[0026] Where: Increase the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period; is the charging power of the electric vehicle charging at the kth charging pile in the charging station at time it0;

[0027] Calculate the frequency regulation capacity reduction of a single electric vehicle:

[0028]

[0029] Where: The lower limit of the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period; is the charging time of the electric vehicle charging at the kth charging pile in the charging station; The minimum power of an electric vehicle charging at the kth charging pile in the charging station;

[0030] The frequency regulation capacity reduction capacity of a single electric vehicle is calculated based on the lower limit of the frequency regulation capacity:

[0031]

[0032] Where: The frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station is reduced in the i-th response period;

[0033] Calculate the frequency regulation capacity and adjustable capacity of a single charging station:

[0034]

[0035] Where: K is the number of electric vehicles under the jurisdiction of the charging station; They are respectively the upward adjustment capacity, downward adjustment capacity, and frequency regulation capacity of the nth charging station under the same cloud.

[0036] Optionally, the method for predicting the frequency regulation load of the subordinate charging stations by the cloud includes: predicting the frequency regulation load of the subordinate charging stations by using the aggregator frequency regulation capacity short-term prediction model, the steps are as follows:

[0037] A1. Call the aggregator's daily frequency modulation capacity historical data stored in the cloud, including the vehicle flow and frequency modulation capacity of the charging stations under its jurisdiction at each time period;

[0038] A2. Preprocess the data, including data cleaning, missing value processing, and outlier detection; divide the processed data into three groups: training set, validation set, and test set;

[0039] A3. Divide the historical data of each period into patches of 96 consecutive data, set charging time, charging piles, and vehicle flow as features to be extracted, and use statistical feature methods to extract them;

[0040] A4. Use the extracted features and the corresponding frequency modulation capacity as target values ​​to train the aggregator frequency modulation capacity short-term prediction model;

[0041] A5. Use the cross-validation method to verify the model and adjust the model's hyperparameters based on the cross-validation results.

[0042] A6. Use the trained aggregator frequency regulation capacity short-term prediction model to predict the frequency regulation load of the charging station.

[0043] Optionally, the method for predicting the frequency regulation load of the subordinate charging stations by the cloud includes: predicting the frequency regulation load of the subordinate charging stations by using an ultra-short-term prediction model of the frequency regulation capacity of the charging stations, the steps are as follows:

[0044] B1. Call the frequency regulation capacity historical data of all response periods of a single charging station in the previous week stored in the cloud, including the increase capacity DP i EVCS，up , reduce capacity DP i EVCS，down ;

[0045] B2. Normalize and divide the above data, and use historical data to initialize the ultra-short-term prediction model of the frequency regulation capacity of the charging station;

[0046] B3. Determine the capacity increase DP of each charging station uploaded by the edge device within the response period i EVCS，up , reduce capacity DP i EVCS ，down The real-time calculation result is the new input data of the ultra-short-term prediction model of the frequency regulation capacity of the charging station, and the network status of the ultra-short-term prediction model of the frequency regulation capacity of the charging station is updated in real time;

[0047] B4. Based on the updated charging station frequency regulation capacity long-term and short-term prediction model, the frequency regulation capacity ultra-short-term prediction is performed, and the output data is the frequency regulation capacity increase capacity of each charging station in the next response period. Reduce capacity

[0048] In some embodiments of the present application, the prediction model may also adopt other time series prediction models to be suitable for different specific application scenarios, which is not limited in the present application.

[0049] Optionally, the decomposition method of the total frequency modulation instruction includes:

[0050] The objective function of cloud command allocation in the i-th response period is established as R t :

[0051]

[0052] Where: is the frequency regulation mileage clearing price for the i-th response period; Allocate adjustment instructions for the nth charging station belonging to the cloud in the ith response period; is the frequency regulation mileage clearing price for the i-th dispatch period; The instruction for the charging station with the same frequency modulation direction as the original instruction to share the frequency modulation direction with the charging station with the opposite frequency modulation direction satisfies the following formula:

[0053]

[0054] Where: is the frequency regulation performance index of the electric vehicle charging station in the i-th scheduling period; δ indicates whether the frequency regulation direction of the charging station is consistent with the total frequency regulation instruction issued by the automatic power generation control master station system, and is calculated as follows:

[0055]

[0056] Optionally, the frequency modulation instruction satisfies a frequency modulation capacity constraint, and the frequency modulation capacity constraint is as follows:

[0057]

[0058] Optionally, the frequency modulation instruction satisfies a user charging demand constraint, and the user charging demand constraint is as follows:

[0059] when At this time, the minimum value instruction that other charging stations in the cloud need to bear is:

[0060]

[0061] In the formula, DP i AGC is the total frequency modulation instruction sent to the cloud during the i-th response period; DP i EVA,dis' For The minimum value instruction that other charging stations in the cloud need to bear in addition to the charging station;

[0062] when When the charging station in the cloud is connected, the instructions that need to be taken are:

[0063]

[0064] In the formula, DP i EVA,dis is the ith response period divided by or The minimum value instruction that other charging stations in the cloud need to bear in addition to the charging station;

[0065] when and When both occur, the command assigned to the charging station is

[0066] Optionally, it is determined whether the charging station with frequency modulation capability can meet the user's charging demand while fully responding to the total frequency modulation instruction issued by the automatic power generation control master station system to the cloud. If the charging station can meet the charging demand while responding to the original instruction, the cloud will respond according to the original instruction; if it cannot meet the charging demand, the cloud will respond according to its maximum frequency modulation capacity. The determination method is as follows:

[0067]

[0068] In the formula, DP i EVA is the actual instruction responded by the cloud in the i-th response period; The frequency regulation capacity of the charging station with frequency regulation capability in the i-th response period is increased by the capacity, and the calculation method is as follows:

[0069]

[0070] The frequency regulation capacity of the charging station with frequency regulation capability is reduced by the following calculation method:

[0071]

[0072] Optionally, on the premise of satisfying the charging needs of users, the frequency modulation instructions allocated by each charging station and each charging pile meet the following conditions:

[0073]

[0074] And the frequency regulation direction of the cloud is the same as the general frequency regulation instruction direction issued by the automatic power generation control master station system:

[0075] DP i AGC DP i EVA >0.

[0076] The beneficial effects of the electric vehicle frequency modulation service system based on the cloud-edge collaborative architecture provided by this application are:

[0077] (1) This application is based on a cloud-edge collaborative control architecture, and the real-time frequency regulation capacity is used to roll out a prediction of the frequency regulation capacity of each charging station in the next response period, which is used as the basis for command decomposition between charging stations, thereby reducing the risk of delay between devices at different levels.

[0078] (2) This application gives priority to the charging needs of electric vehicle users in addition to the expected power consumption on a time scale, ensuring that electric vehicles participate in frequency modulation within the time range expected by users.

[0079] (3) This application ensures that electric vehicles can participate in frequency modulation services while trying their best to meet the needs of the power grid, aggregators, and users through the coordination of cloud, edge devices, and terminal devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0081] Figure 1 A schematic diagram of a cloud-edge collaborative strategy architecture for electric vehicles participating in frequency modulation services provided in an embodiment of the present application;

[0082] Figure 2 A schematic diagram of the decomposition result of the frequency modulation instruction provided in an embodiment of the present application;

[0083] Figure 3 A schematic diagram showing a comparison of system frequency deviations with and without frequency aggregators participating in frequency modulation provided in an embodiment of the present application;

[0084] Figure 4 A schematic diagram of the real-time change results of the electric vehicle power provided in an embodiment of the present application. DETAILED DESCRIPTION

[0085] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0086] See also Figure 1 , the electric vehicle participation frequency modulation service system based on the cloud-edge collaborative architecture provided in the embodiment of the present application is now described.

[0087] An electric vehicle frequency modulation service system based on cloud-edge collaborative architecture includes:

[0088] Terminal device: used to set the user's charging requirements and upload the real-time collected charging pile status information to the edge device;

[0089] Edge devices: used to determine whether electric vehicles can meet the charging requirements set by users during the period of frequency modulation; rolling calculation of the frequency modulation capacity of electric vehicles in the charging station based on the status information of the charging piles and uploading the calculation results to the cloud; receiving the frequency modulation instructions issued by the cloud to the charging station and decomposing them, including: comprehensively adjusting the frequency modulation instructions of each charging pile in the station according to the frequency modulation instructions issued by the cloud, and issuing them to each charging pile;

[0090] Cloud: used to predict the frequency regulation load of the subordinate charging stations according to the calculation results of the edge devices, and upload it to the automatic power generation control master station system; receive the general frequency regulation instruction issued by the automatic power generation control master station system, and decompose it, including: adjusting the frequency regulation instructions of the subordinate charging stations according to the general frequency regulation instruction issued by the automatic power generation control master station system, and issuing it to each charging station;

[0091] Automatic power generation control master station system: sends general frequency regulation instructions to the cloud based on the forecast results of frequency regulation load.

[0092] In some embodiments of the present application, the terminal device is a charging pile.

[0093] In some embodiments of the present application, the edge device is an energy management system in a charging station.

[0094] In some embodiments of the present application, the cloud is a cloud computing center of a load aggregator.

[0095] In some embodiments of the present application, based on the charging demand setting of the electric vehicle user, the edge device will determine whether the electric vehicle can meet the user's demand during the period of frequency modulation. The method for determining whether the charging demand set by the user meets the frequency modulation condition is as follows:

[0096]

[0097] Where: y c =1 and y c =0 means that the charging demand of the user can be met or cannot be met respectively; The maximum power of an electric vehicle charging at the kth charging pile in the charging station; The maximum charging time set for users charging at the kth charging pile in the charging station; α is the minimum threshold coefficient for judging whether the demand can be met, which is generally 15% to 30%; The expected amount of charge to be charged by the user charging at the kth charging pile in the charging station is calculated as follows:

[0098]

[0099] Where: The battery capacity of the electric vehicle charged at the kth charging pile in the charging station; The expected power set by the user who charges at the kth charging pile in the charging station, k is a positive integer greater than or equal to 1 and less than or equal to n, where n is the number of charging piles in the charging station; is the power of the electric vehicle charged at the kth charging pile in the charging station at the time it0; i represents the ith frequency modulation period; it0 is the starting time of the ith response period; where i is a positive integer greater than or equal to 1 and less than or equal to 96, t0 is the frequency modulation capacity evaluation time interval, which is equal to the response time T;

[0100] If the judgment result is yc=0, the user will reset the maximum allowed charging time Expected power Until y c =1.

[0101] In some embodiments of the present application, the method for calculating the frequency modulation capacity of electric vehicles in a charging station is as follows:

[0102] First, calculate the increase in the frequency regulation capacity of a single electric vehicle. The upper limit of the frequency regulation capacity of a single electric vehicle is mainly constrained by the target power set by the user and the battery capacity of the electric vehicle:

[0103]

[0104] Where: The upper limit of the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period;

[0105] Then calculate the frequency regulation capacity increase capacity based on the frequency regulation capacity upper limit:

[0106]

[0107] Where: Increase the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period; is the charging power of the electric vehicle charging at the kth charging pile in the charging station at time it0;

[0108] Then calculate the frequency regulation capacity reduction capacity of a single electric vehicle. The lower limit of the frequency regulation capacity of a single electric vehicle is mainly limited by the maximum charging time allowed by the user:

[0109]

[0110] Where: The lower limit of the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period; is the charging time of the electric vehicle charging at the kth charging pile in the charging station; The minimum power of an electric vehicle charging at the kth charging pile in the charging station;

[0111] Then, the frequency regulation capacity reduction capacity of a single electric vehicle is calculated based on the lower limit of the frequency regulation capacity:

[0112]

[0113] Where: The frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station is reduced in the i-th response period;

[0114] Finally, the frequency regulation capacity and adjustable capacity of a single charging station are calculated. The frequency regulation capacity and adjustable capacity of a single charging station are obtained by aggregating all the individual electric vehicles under its jurisdiction:

[0115]

[0116] Where: K is the number of electric vehicles under the jurisdiction of the charging station; They are respectively the upward adjustment capacity, downward adjustment capacity, and frequency regulation capacity of the nth charging station under the same cloud (aggregator).

[0117] In some embodiments of the present application, the method for the cloud (aggregator) to predict the frequency regulation load of the subordinate charging stations includes: predicting the frequency regulation load of the subordinate charging stations through the aggregator frequency regulation capacity short-term prediction model, the aggregator frequency regulation capacity short-term prediction model is based on the PatchTST prediction model, PatchTST (Patch-based Time Series Forecasting) is a time series prediction model that combines convolutional neural networks (CNN) and long short-term memory networks (LSTM), which is suitable for data with spatiotemporal dependencies. The following are the steps for using the PatchTST prediction model to perform a single-day prediction of the frequency regulation capacity of electric vehicle charging stations:

[0118] A1. Call the aggregator's daily frequency modulation capacity historical data stored in the cloud, including the vehicle flow and frequency modulation capacity of the charging stations under its jurisdiction at each time period;

[0119] A2. Preprocess the data, including data cleaning, missing value processing, and outlier detection. The preprocessing method is a conventional method in the prior art and will not be described here. The processed data is divided into three groups, namely, a training set (60%), a validation set (20%), and a test set (20%).

[0120] A3. Divide the historical data of each time period (i.e., time series data, where the time interval of historical data is 15 minutes, and there are 96 15-minute intervals in a day) into patches with 96 continuous data as units, set charging time, charging piles, and vehicle flow as features to be extracted, and use statistical feature methods to extract them;

[0121] A4. Use the extracted features and the corresponding frequency modulation capacity as the target value to train the PatchTST prediction model, with 100 training rounds;

[0122] A5. Use the cross-validation method to verify the model, and adjust the model's hyperparameters based on the cross-validation results to improve the model's performance.

[0123] A6. The trained PatchTST prediction model is the aggregator's frequency regulation capacity short-term prediction model. The frequency regulation load of the charging station is predicted using the aggregator's frequency regulation capacity short-term prediction model.

[0124] In some embodiments of the present application, in order to alleviate the delay problem between devices at different levels, the cloud can use online learning to predict the frequency modulation capacity of each charging station in the next response phase in real time, which is used as the basis for the allocation of instructions between charging stations. For the frequency modulation capacity of a single charging station, an ultra-short-term prediction model of the frequency modulation capacity of the charging station based on the LSTM (long short-term memory network) model is used for prediction. The LSTM model has a strong memory capacity and can better capture the dependencies in the time series. The specific prediction method is as follows:

[0125] B1. Call the frequency regulation capacity historical data of all response periods of a single charging station in the previous week stored in the cloud, including the increase capacity DP i EVCS，up , reduce capacity DP i EVCS，down ;

[0126] B2. Normalize and divide the above data, and use historical data to initialize the LSTM prediction model;

[0127] B3. Determine the capacity increase DP of each charging station uploaded by the edge device within the response period i EVCS，up , reduce capacity DP i EVCS ，down The real-time calculation result is the new input data of the LSTM prediction model, and the network status of the LSTM prediction model is updated in real time;

[0128] B4. The updated LSTM prediction model is the ultra-short-term prediction model for the frequency regulation capacity of the charging station. The ultra-short-term prediction model is used to perform the ultra-short-term prediction of the frequency regulation capacity. The output data is the increase capacity of the frequency regulation capacity of each charging station in the next response period. Reduce capacity

[0129] The benefits of the aggregator participating in the frequency regulation auxiliary service include frequency regulation mileage compensation and frequency regulation capacity compensation. The objective function of the cloud (load aggregator) instruction allocation in the i-th response period is established as follows:

[0130] The objective function of cloud command allocation in the i-th response period is established as Rt :

[0131]

[0132] Where: is the frequency regulation mileage clearing price for the i-th response period; Allocate adjustment instructions to the nth charging station belonging to the cloud (aggregator) in the ith response period; is the frequency regulation mileage clearing price for the i-th dispatch period; The instruction for the charging station with the same frequency modulation direction as the original instruction to share the frequency modulation direction with the charging station with the opposite frequency modulation direction satisfies the following formula:

[0133]

[0134] Where: is the frequency regulation performance index of the electric vehicle charging station in the i-th scheduling period; δ indicates whether the frequency regulation direction of the charging station is consistent with the total frequency regulation instruction issued by the automatic power generation control master station system, and the calculation method is as follows:

[0135]

[0136] In some embodiments of the present application, the frequency modulation instruction is limited within the frequency modulation capacity range, that is, the frequency modulation instruction satisfies the frequency modulation capacity constraint, and the frequency modulation capacity constraint is as follows:

[0137]

[0138]

[0139] Considering the user's charging needs in terms of power and time, two situations need to be given priority before allocating other instructions.

[0140] when When , it means that the electric vehicles in the charging station are about to complete charging. At this time, the power needs to be reduced to avoid overcharging, so the charging station is allocated a downward adjustment command. At this time, the minimum value command that other charging stations in the cloud (aggregator) need to bear is:

[0141]

[0142] In the formula, DP i AGC is the total frequency modulation instruction sent to the cloud during the i-th response period; DP i EVA,dis' For The minimum value instruction that other charging stations in the cloud need to bear in addition to the charging station;

[0143] when , it means that the longest charging time allowed for users in this charging station is about to end, and the power needs to be increased to meet user needs. Therefore, the charging station is assigned an upward adjustment command. The commands that other charging stations in the cloud (aggregator) need to bear are:

[0144]

[0145] In the formula, DP i EVA,dis is the ith response period divided by or The minimum value instruction that other charging stations in the cloud need to bear in addition to the charging station;

[0146] when and When both occur, the command assigned to the charging station is This ensures that the electric vehicle is fully charged during this period without overcharging.

[0147] Based on the above calculations, it is determined whether the charging station with frequency modulation capability can meet the user's charging needs while fully responding to the total frequency modulation command issued by the automatic power generation control master station system to the cloud. If the charging station can meet the charging needs while responding to the original command, the cloud (aggregator) will respond according to the original command; if it cannot meet the charging needs, the cloud (aggregator) will respond according to its maximum frequency modulation capacity. The judgment method is as follows:

[0148]

[0149] In the formula, DP i EVA is the actual instruction responded by the cloud in the i-th response period; The frequency regulation capacity of the charging station with frequency regulation capability in the i-th response period is increased by the capacity, and the calculation method is as follows:

[0150]

[0151] The frequency regulation capacity of the charging station with frequency regulation capability is reduced by the following calculation method:

[0152]

[0153] The allocation of frequency modulation commands between electric vehicles must also meet the above constraints. The situation is the same and will not be repeated here.

[0154] The cloud (aggregator) should fully respond to the frequency modulation needs of the power grid while meeting the charging needs of users. Therefore, the frequency modulation instructions allocated to each charging station and each charging pile must meet the following conditions:

[0155]

[0156] In addition, the frequency regulation direction of the cloud is the same as the general frequency regulation instruction direction issued by the automatic power generation control master station system, that is:

[0157] DP i AGC DP i EVA >0.

[0158] The electric vehicle frequency modulation service system based on the cloud-edge collaborative architecture provided in this application has developed a frequency modulation capacity evaluation and prediction method and a frequency modulation capacity decomposition optimization strategy based on the cloud-edge collaborative control architecture, so as to achieve coordinated electric vehicle participation in frequency modulation services while ensuring the charging needs of users. Compared with the prior art, the advantage of this system is that it uses the frequency modulation capacity of the next response period predicted in real time as the basis for command decomposition between charging stations, reducing the risk of delays between devices at different levels; it also gives priority to the charging needs of electric vehicle users on a time scale in addition to the expected power, ensuring that electric vehicles participate in frequency modulation within the time range expected by users; at the same time, the command decomposition strategy formulated with the goal of maximizing the benefits of aggregators can be better applied to the power auxiliary market.

[0159] Example

[0160] The whole day is divided into 96 response periods, each of which lasts 15 minutes. The frequency regulation instructions issued by the power grid refer to the 31-day ACE data of the PJM market, simulate the process of prediction, declaration, clearing, and response, and the final response instructions are converted according to the frequency regulation capacity predicted by the aggregator. It is assumed that the aggregator has seven charging stations under its jurisdiction, and each charging station has seven charging piles. The probability distribution of data such as electric vehicle operating parameters and charging station traffic is shown in Tables 1 and 2. The charging behavior of users is simulated using the Monte Carlo method.

[0161] Table 1 Electric vehicle operating parameters

[0162]

[0163] Table 2 Traffic flow at charging stations

[0164]

[0165] 1) Frequency regulation capacity prediction results

[0166] In the short-term prediction of frequency modulation capacity of aggregators, PatchTST is compared with N-BEATS, N-HiTS, and LSTM models using the same frequency modulation capacity dataset. The results are evaluated by root mean square error, as shown in Table 3:

[0167] Table 3 Aggregator frequency modulation capacity prediction results

[0168]

[0169] As can be seen from Table 3, the prediction error of the PatchTST model is the smallest among the four models, and the root mean square error is also small, indicating that the PatchTST model has the conditions to provide a basis for declaration for aggregators in frequency modulation capacity prediction.

[0170] In the ultra-short-term prediction of the frequency regulation capacity of the charging station, the above model is also used for comparison, and the results are shown in Table 4:

[0171] Table 4 Prediction results of charging station frequency regulation capacity

[0172]

[0173] As can be seen from Table 4, the LSTM model has a clear advantage in prediction error. In addition, in the simulation environment of this paper, the trained LSTM model takes an average of only 7.8 ms to perform ultra-short-term prediction, indicating that it can assist in the rapid decomposition of instructions between charging stations.

[0174] 2) Frequency modulation instruction decomposition results

[0175] The command decomposition results of the charging station 1 under the control of the aggregator and the electric vehicles charging at the charging pile 1 under the control of the charging station 1 on a certain operation day are as follows: Figure 2 As shown, Figure 2 (a) is the command decomposition result of the charging station. Figure 2 (b) is the command decomposition result of electric vehicles, which shows that the frequency modulation command can be effectively decomposed within the frequency modulation capacity range based on edge computing.

[0176] 3) FM effect analysis

[0177] According to the actual response results of the aggregators during the six-day response period, Figure 3 The comparison results of system frequency deviation with and without the participation of aggregators are shown. Although the aggregator does not have the ability to control the frequency deviation to 0Hz in some periods due to factors such as prediction errors, it can be found that the participation of aggregators effectively reduces the system frequency deviation in each period. In some periods, the aggregator can ideally and continuously respond to actual instructions, indicating that the participation of aggregators in frequency modulation can effectively reduce the system frequency deviation.

[0178] 4) User needs analysis

[0179] The real-time results of the electric vehicle power charged at charging pile 1 under charging station 1 on a certain operation day are as follows: Figure 4 As shown, all electric vehicles can participate in frequency regulation while ensuring users' needs for target power and maximum allowed charging time.

[0180] 5) Analysis of aggregator benefits

[0181] Since the frequency regulation resources under the control of the aggregator in the example are relatively small and insufficient to affect the market clearing price, the frequency regulation mileage compensation price D for each period of the operation day in the example simulation is specified. i M =80 and frequency regulation capacity compensation price D i C = 12. Based on the set price, the average frequency regulation mileage compensation that the aggregator can obtain in a single day in the simulation is about 297,092.88 yuan, and the average frequency regulation capacity compensation is about 379,189.67 yuan, which reflects the economic benefits of the aggregator aggregating and coordinating electric vehicles to participate in frequency regulation services.

[0182] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. An electric vehicle frequency modulation service system based on cloud-edge collaborative architecture, characterized in that: include: Terminal device: used to set the user's charging requirements and upload the real-time collected charging pile status information to the edge device; Edge devices: used to determine whether electric vehicles can meet the charging requirements set by users during the period of frequency modulation; rolling calculation of the frequency modulation capacity of electric vehicles in the charging station based on the status information of the charging piles and uploading the calculation results to the cloud; Receive the frequency modulation command issued by the cloud to the charging station and decompose it, including: comprehensively adjust the frequency modulation command of each charging pile in the station according to the frequency modulation command issued by the cloud, and send it to each charging pile; Cloud: used to predict the frequency regulation load of the subordinate charging stations according to the calculation results of the edge devices, and upload it to the automatic power generation control master station system; receive the general frequency regulation instruction issued by the automatic power generation control master station system, and decompose it, including: adjusting the frequency regulation instructions of the subordinate charging stations according to the general frequency regulation instruction issued by the automatic power generation control master station system, and issuing it to each charging station; Automatic power generation control master station system: sends general frequency regulation instructions to the cloud based on the forecast results of frequency regulation load.

2. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 1, characterized in that: The method for judging whether the charging demand set by the user meets the frequency modulation condition is as follows: Where: y c =1 and y c =0 means that the charging demand of the user can be met or cannot be met respectively; The maximum power of an electric vehicle charging at the kth charging pile in the charging station; The maximum charging time set for users charging at the kth charging pile in the charging station; α is the minimum threshold coefficient for judging whether the demand can be met; The expected amount of charge to be charged by the user charging at the kth charging pile in the charging station is calculated as follows: Where: The battery capacity of the electric vehicle charged at the kth charging pile in the charging station; The expected power set for the user charging at the kth charging pile in the charging station; is the power of the electric vehicle charged at the kth charging pile in the charging station at the time it0; i represents the ith frequency modulation period; it0 is the starting time of the ith response period; where i is a positive integer greater than or equal to 1 and less than or equal to 96, and t0 is the frequency modulation capacity evaluation time interval; If the judgment result is y c = 0, the user will reset the maximum allowed charging time Expected power Until y c =1.

3. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 1, characterized in that: The method for calculating the frequency regulation capacity of electric vehicles in charging stations is as follows: Calculate the increase in frequency regulation capacity of a single electric vehicle: Where: The upper limit of the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period; Calculate the frequency regulation capacity increase capacity based on the frequency regulation capacity upper limit: Where: Increase the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period; is the charging power of the electric vehicle charging at the kth charging pile in the charging station at time it0; Calculate the frequency regulation capacity reduction of a single electric vehicle: Where: The lower limit of the frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station during the i-th response period; is the charging time of the electric vehicle charging at the kth charging pile in the charging station; The minimum power of an electric vehicle charging at the kth charging pile in the charging station; The frequency regulation capacity reduction capacity of a single electric vehicle is calculated based on the lower limit of the frequency regulation capacity: Where: The frequency regulation capacity of the electric vehicle charging at the kth charging pile in the charging station is reduced in the i-th response period; Calculate the frequency regulation capacity and adjustable capacity of a single charging station: Where: K is the number of electric vehicles under the jurisdiction of the charging station; They are respectively the upward adjustment capacity, downward adjustment capacity, and frequency regulation capacity of the nth charging station under the same cloud.

4. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 3 is characterized in that: The method for predicting the frequency regulation load of the subordinate charging stations in the cloud includes: predicting the frequency regulation load of the subordinate charging stations through the aggregator frequency regulation capacity short-term prediction model, and the steps are as follows: A1. Call the aggregator's daily frequency modulation capacity historical data stored in the cloud, including the vehicle flow and frequency modulation capacity of the charging stations under its jurisdiction at each time period; A2. Preprocess the data, including data cleaning, missing value processing, and outlier detection; divide the processed data into three groups: training set, validation set, and test set; A3. Divide the historical data of each period into patches of 96 consecutive data, set charging time, charging piles, and vehicle flow as features to be extracted, and use statistical feature methods to extract them; A4. Use the extracted features and the corresponding frequency modulation capacity as target values ​​to train the aggregator frequency modulation capacity short-term prediction model; A5. Use the cross-validation method to verify the model and adjust the model's hyperparameters based on the cross-validation results. A6. Use the trained aggregator frequency regulation capacity short-term prediction model to predict the frequency regulation load of the charging station.

5. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 3, characterized in that: The method for predicting the frequency regulation load of the subordinate charging stations in the cloud includes: predicting the frequency regulation load of the subordinate charging stations through the ultra-short-term prediction model of the frequency regulation capacity of the charging stations, and the steps are as follows: B1. Call the historical data of frequency regulation capacity of a single charging station in all response periods in the previous week stored in the cloud, including the upward regulation capacity Reduce capacity B2. Normalize and divide the above data, and use historical data to initialize the ultra-short-term prediction model of the frequency regulation capacity of the charging station; B3. Determine the increased capacity of each charging station uploaded by the edge device within the response period Reduce capacity The real-time calculation result is the new input data of the ultra-short-term prediction model of the frequency regulation capacity of the charging station, and the network status of the ultra-short-term prediction model of the frequency regulation capacity of the charging station is updated in real time; B4. Based on the updated charging station frequency regulation capacity long-term and short-term prediction model, the frequency regulation capacity ultra-short-term prediction is performed, and the output data is the frequency regulation capacity increase capacity of each charging station in the next response period. Reduce capacity 6. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 3, characterized in that: The decomposition method of the total frequency modulation instruction includes: The objective function of cloud command allocation in the i-th response period is established as R t : Where: is the frequency regulation mileage clearing price for the i-th response period; Allocate adjustment instructions for the nth charging station belonging to the cloud in the ith response period; is the frequency regulation mileage clearing price for the i-th dispatch period; The instruction for the charging station with the same frequency modulation direction as the original instruction to share the frequency modulation direction with the charging station with the opposite frequency modulation direction satisfies the following formula: Where: is the frequency regulation performance index of the electric vehicle charging station in the i-th scheduling period; δ indicates whether the frequency regulation direction of the charging station is consistent with the total frequency regulation instruction issued by the automatic power generation control master station system, and the calculation method is as follows:

7. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 6, characterized in that: The frequency modulation instruction satisfies the frequency modulation capacity constraint, which is as follows:

8. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 6, characterized in that: The frequency modulation instruction meets the user charging demand constraints, which are as follows: when At this time, the minimum value instruction that other charging stations in the cloud need to bear is: In the formula, is the total frequency modulation instruction sent to the cloud in the i-th response period; For The minimum value instruction that other charging stations in the cloud need to bear in addition to the charging station; when When the charging station in the cloud is connected, the instructions that need to be taken are: In the formula, is the ith response period divided by or The minimum value instruction that other charging stations in the cloud need to bear in addition to the charging station; when and When both occur, the command assigned to the charging station is 9. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 8, characterized in that: Determine whether the charging station with frequency modulation capability can meet the user's charging needs while fully responding to the total frequency modulation command issued by the automatic power generation control master station system to the cloud. If the charging station can meet the charging needs while responding to the original command, the cloud will respond according to the original command; if it cannot meet the charging needs, the cloud will respond according to its maximum frequency modulation capacity. The judgment method is as follows: In the formula, is the actual instruction responded by the cloud in the i-th response period; The frequency regulation capacity of the charging station with frequency regulation capability in the i-th response period is increased by the capacity, and the calculation method is as follows: The frequency regulation capacity of the charging station with frequency regulation capability is reduced by the following calculation method:

10. The electric vehicle frequency modulation service system based on cloud-edge collaborative architecture as claimed in claim 9, characterized in that: On the premise of meeting the charging needs of users, the frequency modulation instructions allocated by each charging station and each charging pile meet the following conditions: And the frequency regulation direction of the cloud is the same as the general frequency regulation instruction direction issued by the automatic power generation control master station system:

Images