Charging load regulation and control method and system and terminal equipment

Abstract

The invention provides a charging load regulation and control method and system, terminal equipment and a medium, and belongs to the field of charging load regulation and control, and the method comprises the steps: obtaining historical operation data of a charging station, including a charging event sample of each charging user during each charging; based on the charging event sample, extracting a quantity feature, a charging starting time feature and a charging duration feature of the charging users; constructing a behavior track of the charging user according to the charging starting time feature and the charging duration feature; carrying out Monte Carlo simulation based on the quantity characteristics and the behavior tracks, and constructing a plurality of charging load scenes; under a preset constraint condition, calling a preset solver to solve each charging load scene, and generating a load prediction curve of each charging scene; and the charging parameters of the charging station are regulated and controlled according to the load prediction curves. According to the invention, the deviation between the charging load and the actual load scene is effectively reduced, and the authenticity of load prediction is optimized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of charging load regulation, and in particular to a charging load regulation method, system and terminal device. BACKGROUND

[0002] With the rapid popularization of new energy vehicles, public and private charging facilities are accelerating the expansion, and charging load presents obvious concentration and randomness in time and space. In addition to considering the vehicle power, the prediction of charging load usually considers factors such as electricity price, user commuting, travel, and weather. Therefore, accurate charging prediction and reasonable scheduling to a certain extent can not only improve user satisfaction, but also ensure the efficient and stable operation of the power system.

[0003] In the aspect of charging load prediction, the existing methods mostly rely on historical statistics, time series or machine learning models for prediction. For example, some technical solutions assume that user behavior is stable, and generate a deterministic load curve through historical data fitting. However, this technical solution cannot cover the multi-factor random scenario of user number fluctuation and random charging time in actual operation, and the charging load prediction result based on this technical solution will deviate greatly from the actual load scenario, and the authenticity is low. SUMMARY

[0004] The present application provides a charging load regulation method, system and terminal device to solve the technical problem of how to reduce the deviation between charging load and actual load scenario.

[0005] To solve the above technical problem, the present application provides a charging load regulation method, comprising: obtaining historical operation data of a charging station; the historical operation data includes a charging event sample of each charging user for each charging; based on the charging event sample, extracting the number of charging users, the charging start time feature and the charging duration feature; According to the charging start time feature and the charging duration feature, the behavior trajectory of the charging user is constructed, and the Monte Carlo simulation is performed based on the number of features and the behavior trajectory to construct multiple charging load scenarios; Under the preset constraint condition, a preset solver is called to solve each charging load scenario to generate a load prediction curve for each charging scenario; and the charging parameters of the charging station are regulated according to the load prediction curve.

[0006] As a preferred solution, the charging station includes a plurality of charging piles; Before obtaining the historical operation data of the charging station, it includes: acquiring the running state data of the plurality of charging piles through a preset sensor; extracting operation records of each charging pile respectively based on the operation state data; According to the operation records, a charging event is generated each time the charging behavior of a charging user is monitored, and the generated charging events are constructed as the charging event samples.

[0007] As a preferred solution, the number of charging users, the charging start time feature and the charging duration feature are extracted based on the charging event samples, including: dividing each charging event sample according to time to obtain sample data of multiple discrete time periods; extracting user feature data from the sample data of the multiple discrete time periods by taking a preset data field as a search key, wherein the user feature data is represented in a specific data type; extracting the number of charging users, the charging start time feature and the charging duration feature from the user feature data.

[0008] As a preferred solution, the number of charging users, the charging start time feature and the charging duration feature are extracted from the user feature data, including: constructing a Poisson distribution function of the number of charging users based on the user feature data, and determining the Poisson distribution function as the number feature; constructing a probability distribution model of the charging start time of the charging user based on the user feature data, and determining the probability distribution model as the charging start time feature; fitting a probability density function of the charging duration of the charging user based on the user feature data, and determining the probability density function as the charging duration feature.

[0009] As a preferred solution, the behavior trajectory of the charging user is constructed according to the charging start time feature and the charging duration feature, and a plurality of charging load scenarios are constructed based on the number feature and the behavior trajectory through Monte Carlo simulation, including: randomly extracting the charging start time of the charging user according to the probability distribution model, and randomly extracting the charging duration of the charging user according to the charging duration feature; constructing the behavior trajectory of the charging user of each charging load scenario according to the extracted charging start time and the extracted charging duration; randomly generating the number of charging users of each charging load scenario based on the number feature; constructing a plurality of charging load scenarios based on the constructed behavior trajectory and the randomly generated number of charging users through Monte Carlo simulation.

[0010] Preferably, the constructing the probability distribution model of the charging start time of the charging user based on the user feature data comprises: calculating a first probability of the user starting charging based on the user feature data and a time dimension; obtaining a preset price sensitivity parameter; and calculating a second probability of the user starting charging based on the user feature data, the price sensitivity parameter and an electricity price dimension; calculating a joint probability distribution based on the first probability and the second probability to obtain the probability distribution model.

[0011] Preferably, the user feature data comprises a mean value of the charging duration of the sample data; and the fitting the probability density function of the charging duration of the charging user based on the user feature data comprises: calculating a sample standard deviation according to the mean value and each charging duration in the sample data; calculating a bandwidth by the Silverma rule based on the standard deviation and a sample quantity; performing kernel density estimation according to the sample quantity, the bandwidth and each charging duration in the sample data to obtain a Gaussian kernel function; determining the Gaussian kernel function as the probability density function.

[0012] Preferably, the calling a preset solver to solve each charging load scenario under a preset constraint to generate a load prediction curve of each charging scenario comprises: under preset charging pile power constraints, charging service fee constraints and user satisfaction constraints, using a Cplex solver or a Gurobi solver to solve each charging load scenario to generate a load prediction curve of each charging scenario.

[0013] Correspondingly, the application also provides a charging load regulation system, comprising a historical data acquisition module, a feature extraction module, a simulation scenario construction module and a regulation module; wherein, the historical data acquisition module is configured to acquire historical operation data of a charging station; the historical operation data comprises charging event samples of each charging user each time; the feature extraction module is configured to extract a quantity feature, a charging start time feature and a charging duration feature of a charging user based on the charging event samples; the simulation scenario construction module is configured to construct a behavior trajectory of a charging user according to the charging start time feature and the charging duration feature; and perform Monte Carlo simulation based on the quantity feature and the behavior trajectory to construct a plurality of charging load scenarios; The regulation module is configured to call a preset solver to solve each charging load scenario under preset constraint conditions, to generate a load prediction curve of each charging scenario, and to regulate and control charging parameters of the charging station according to the load prediction curve.

[0014] As a preferred solution, the charging station comprises a plurality of charging piles, and the regulation system comprises a charging event generation module configured to, before the historical operation data of the charging station is acquired: acquire running state data of the plurality of charging piles through a preset sensor; extract running records of each charging pile based on the running state data; generate a charging event each time a charging behavior of a charging user is monitored according to the running records, and construct the generated charging event as the charging event sample.

[0015] As a preferred solution, the feature extraction module extracts a quantity feature, a charging start time feature, and a charging duration feature of a charging user based on the charging event sample, including: the feature extraction module divides each charging event sample according to time to obtain sample data of a plurality of discrete time periods;

[0016] extracts user feature data from the sample data of the plurality of discrete time periods by taking a preset data field as a search key, wherein the user feature data is represented in a specific data type; extracts the quantity feature, the charging start time feature, and the charging duration feature of the charging user from the user feature data. As a preferred solution, the feature extraction module extracts a quantity feature, a charging start time feature, and a charging duration feature of a charging user from the user feature data, including: the feature extraction module constructs a Poisson distribution function of the quantity of the charging user based on the user feature data, and determines the Poisson distribution function as the quantity feature;

[0017] constructs a probability distribution model of the charging start time of the charging user based on the user feature data, and determines the probability distribution model as the charging start time feature; fits a probability density function of the charging duration of the charging user based on the user feature data, and determines the probability density function as the charging duration feature. As a preferred solution, the simulation scenario construction module constructs a behavior trajectory of a charging user according to the charging start time feature and the charging duration feature, and performs Monte Carlo simulation based on the quantity feature and the behavior trajectory to construct a plurality of charging load scenarios, including: The simulation scenario construction module randomly extracts a charging start time of the charging user according to the probability distribution model; and randomly extracts a charging duration of the charging user according to the charging duration feature; According to the extracted charging start time and the extracted charging duration, a behavior trajectory of the charging user of each charging load scenario is constructed; Based on the quantity feature, the number of charging users of each charging load scenario is randomly generated; Based on the constructed behavior trajectory and the randomly generated number of charging users, Monte Carlo simulation is performed to construct a plurality of charging load scenarios.

[0018] As a preferred solution, the feature extraction module constructs a probability distribution model of the charging start time of the charging user based on the user feature data, including: The feature extraction module calculates a first probability of the user starting charging based on the user feature data and a time dimension; A preset price sensitivity parameter is obtained; and a second probability of the user starting charging is calculated based on the user feature data, the price sensitivity parameter and a price dimension; Based on the first probability and the second probability, a joint probability distribution is calculated to obtain the probability distribution model.

[0019] As a preferred solution, the user feature data includes a mean value of the charging duration of the sample data; and the feature extraction module fits a probability density function of the charging duration of the charging user based on the user feature data, including: The feature extraction module calculates a sample standard deviation based on the mean value and each charging duration in the sample data; Based on the standard deviation and the sample quantity, a bandwidth is calculated by the Silverma rule; According to the sample quantity, the bandwidth and each charging duration in the sample data, kernel density estimation is performed to obtain a Gaussian kernel function; The Gaussian kernel function is determined as the probability density function.

[0020] As a preferred solution, the regulation module calls a preset solver to solve each charging load scenario respectively under a preset constraint condition to generate a load prediction curve of each charging scenario, including: The regulation module adopts a Cplex solver or a Gurobi solver to solve each charging load scenario respectively under a preset charging pile power constraint, a charging service fee constraint and a user satisfaction constraint to generate a load prediction curve of each charging scenario.

[0021] Correspondingly, the application further provides a terminal device, comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the charging load regulation method when executing the computer program.

[0022] Compared with the prior art, the application has the following beneficial effects: The application provides a charging load regulation method, system, terminal device and medium. By extracting features from charging event samples, number features, charging start time features and electrical duration features are obtained, and then a plurality of charging load scenarios are generated through Monte Carlo simulation. Compared with the prior art, the plurality of charging load scenarios can cover scenarios with random factors such as user number fluctuation, charging start time and charging duration in actual operation, that is, the differences in charging time caused by user behavior differences are considered, the one-sidedness of the charging load prediction of the prior art solution is solved to a certain extent, the deviation between the charging load and the actual load scenario is effectively reduced, and the authenticity of the load prediction is optimized.

[0023] In addition, by calling a preset solver to solve each charging load scenario, a load prediction curve (i.e. a plurality of load prediction curves) of each charging load scenario is generated. Compared with the prior art technical solution of generating only a single curve (a single load curve is often based on the assumption of stable user behavior and only reflects the average law of the load, and cannot cover extreme load or sudden fluctuation scenarios), the load fluctuation under different user behavior trajectories can be covered, the random fluctuation characteristics of the load are reflected, and the prediction result is further closer to the actual scenario of charging operation.

[0024] Finally, the charging parameters of the charging station are regulated according to each load prediction curve, which can ensure the efficient and stable operation of the charging station on the basis of accurate and real load prediction, and reduce the risk caused by extreme load or sudden large fluctuations. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A flowchart of an embodiment of the charging load regulation method provided by the application.

[0026] Figure 2 A flowchart of a preferred embodiment of an embodiment of the charging load regulation method provided by the application.

[0027] Figure 3 A columnar diagram of an example of the total number of charging users provided by the application.

[0028] Figure 4 A columnar diagram of an example of the Poisson distribution corresponding to the total number of charging users provided by the application.

[0029] Figure 5 A probability distribution diagram of an example of the first probability provided for the present application.

[0030] Figure 6 A probability distribution diagram of an example of the second probability provided for the present application.

[0031] Figure 7 A probability distribution diagram of an example of the joint probability distribution provided for the present application.

[0032] Figure 8 A diagram of an embodiment of the Gaussian kernel function of the charging duration provided for the present application.

[0033] Figure 9 A diagram of an example of the charging behavior trajectory of the charging load scenario s provided for the present application.

[0034] Figure 10 A diagram of an example of the charging power of the charging load scenario s provided for the present application.

[0035] Figure 11 A structural diagram of an embodiment of the regulation system of the charging load provided for the present application. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0037] Embodiment One Please refer to Figure 1 , Figure 1 A flow diagram of an embodiment of the regulation method of the charging load provided for the present application. Figure 1 The illustrated embodiment includes steps S101 to S104; each step is described in detail as follows: Step S101, obtaining historical operation data of charging stations.

[0038] In this step, one or more charging stations can be provided in each region, and each charging station can be provided with multiple charging piles. This step collects the historical operation data of one or more charging stations in the region, which is used as the data basis for subsequent feature extraction, scenario simulation and load prediction.

[0039] In the embodiment, each charging user can generate a kind of "arrival event" when his electric vehicle accesses the charging pile and charges each time, and the historical operation data includes the charging event of each charging user each time as a sample, hereinafter referred to as a charging event sample.

[0040] Exemplarily, before the step S101 of obtaining the historical operation data of the charging station, the step includes: collecting the operation state data of the plurality of charging piles through a preset sensor; extracting the operation record of each charging pile based on the operation state data; and generating a charging event each time the charging behavior of a charging user is monitored according to the operation record, and constructing the generated charging event as the charging event sample.

[0041] In the embodiment, one or more sensors can be pre-set on each charging pile, and the types of the sensors include but are not limited to voltage sensors or current sensors. The purpose of setting the sensors is to collect the operation state data of the charging pile, and then analyze the charging behavior of the charging user.

[0042] In step S102, the number of charging users, the charging start time feature and the charging duration feature are extracted based on the charging event sample.

[0043] In this step, by analyzing the charging event sample, the number of charging users can be analyzed, and the behavior rule of the user can be analyzed to some extent, such as the charging start time feature and the charging duration feature.

[0044] In some preferred embodiments, as shown in Figure 2 The step S102 of extracting the number of charging users, the charging start time feature and the charging duration feature based on the charging event sample includes steps S201 to S203, and each step is described as follows: In step S201, the charging event sample is divided according to time to obtain sample data of a plurality of discrete time periods. In step S202, the user feature data is extracted from the sample data of the plurality of discrete time periods by taking a preset data field as a retrieval key, and the user feature data is represented by a specific data type. In step S203, the number of charging users, the charging start time feature and the charging duration feature are extracted from the user feature data.

[0045] In the embodiment, the step S201 of dividing the charging event sample according to time can be specifically performed on M charging piles and D days of data, and exemplarily, T discrete time periods can be divided each day.

[0046] The user feature data is extracted from the sample data of the plurality of discrete time periods by taking the preset data field as a search key, and the user feature data includes but is not limited to time, time period, number of users charging in the time period, real-time electricity price, etc., and one example is shown in Table 1: Table 1: Example of user feature data In step S202, the user feature data can be extracted from the sample data of the plurality of discrete time periods by taking the preset data field (corresponding to the field name in the table) as a search key, and the user feature data is represented in a specific data type, such as time period number, number of users charging in the time period, etc., and the number of users charging, charging start time and charging duration are extracted.

[0047] In some embodiments, the charging price C cg (t) can be divided into B non-overlapping intervals (where b refers to the b-th interval), such as For each price interval, the total number of users charging in D days falling into the interval is denoted as N b , and the center value of the interval can be represented as: ; In addition, the charging duration data of all users are arranged in ascending order, and the charging duration data of all users in D days can be obtained as , …, , where m refers to the m-th charging duration or the total number of samples of charging duration.

[0048] Preferably, the extracting of the number of users charging, the charging start time and the charging duration from the user feature data comprises: constructing a Poisson distribution function of the number of users charging based on the user feature data, and determining the Poisson distribution function as the number characteristic; constructing a probability distribution model of the charging start time of the users charging based on the user feature data, and determining the probability distribution model as the charging start time characteristic; fitting a probability density function of the charging duration of the users charging based on the user feature data, and determining the probability density function as the charging duration characteristic.

[0049] In terms of the number of users charging, an example is shown in Figure 3 and 4 In actual operation scenarios, the number of users entering the charging station each day has strong randomness, and the behavior of users accessing charging each time or each day is a "charging event".

[0050] This example uses the Poisson distribution function to represent the total number of daily charging users N. s Perform modeling and obtain The calculated Poisson distribution function is used as a characteristic of the number of charging users. It should be noted that... This refers to the total number of users who charge their devices on day D.

[0051] In some other preferred embodiments, two main factors are considered in the extraction of charging start time features: one is that the user's consideration of starting charging is time-driven, and the other is that the user's consideration of starting charging is electricity price-driven.

[0052] For example, constructing a probability distribution model of the charging start time of the charging user based on the user feature data includes: calculating a first probability that the user starts charging based on the user feature data and the time dimension; obtaining a preset price-sensitive parameter, and calculating a second probability that the user starts charging based on the user feature data, the price-sensitive parameter, and the electricity price dimension; and calculating a joint probability distribution based on the first probability and the second probability to obtain the probability distribution model.

[0053] For the time-driven feature extraction mentioned above, such as Figure 5 As shown, the charging behavior of electric vehicle users is often influenced by their daily travel patterns. For example, some users start charging before or after commuting or after returning home in the evening. This behavior has a certain degree of time concentration and regularity.

[0054] The probability that the user starts charging within time t (i.e., the first probability) is: ; Among them, P time Let t be the first probability, T be the total number of time periods, D be the total number of days, and t be the first probability. start =t indicates that the first probability is calculated for the t-th time period, N d,t This refers to the total number of users charging in the t-th time period.

[0055] For the above-mentioned feature extraction driven by electricity prices, such as Figure 6 As shown, electric vehicle users' choice of charging time is driven not only by their daily travel patterns but also by the charging costs (i.e., the sum of electricity price and service fee) at different times. To characterize users' time selection behavior based on price factors, this embodiment can use a logit model to model users' price sensitivity. Let the price sensitivity parameter be... Then the user in the first The probability (second probability) of choosing to start charging due to price factors within a given time period is: ; wherein P cost is the second probability, ) denotes the charging price of the tth time period.

[0056] wherein the price sensitivity parameter is a preset value. The value thereof is determined as follows: First, an empirical probability distribution is obtained according to the charging event samples: ; wherein N b denotes the total number of charging users in the bth time period or time interval in the charging event samples, and B denotes the total number of time periods or time intervals in the charging event samples; The least square method is used to fit the difference between the model distribution and the empirical distribution, and a loss function is constructed: ; The optimal price sensitivity parameter is: .

[0057] Further, as shown in Figure 7 , in order to comprehensively consider the joint influence of time driving and price driving, the joint probability distribution can be calculated according to the following formula, and the probability distribution model is obtained: ; wherein P denotes the joint probability distribution.

[0058] In some preferred embodiments, as shown in Figure 8 , the user feature data includes the mean of the charging duration of the sample data; and the fitting of the probability density function of the charging duration of the charging user based on the user feature data includes: calculating the sample standard deviation according to the mean and each charging duration in the sample data; calculating the bandwidth by the Silverma rule based on the standard deviation and the sample quantity; performing kernel density estimation according to the sample quantity, the bandwidth and each charging duration in the sample data to obtain a Gaussian kernel function; and determining the Gaussian kernel function as the probability density function.

[0059] The calculation formula of the mean is: ; wherein m denotes the total number of samples of the charging duration, k denotes the sample of the kth charging duration, denotes the mean; Further, the standard deviation is calculated as: ; Based on the standard deviation combined with the sample number, the bandwidth h is calculated by Silverma rule: ; The calculated bandwidth h can be used to adjust the smoothing degree of the kernel density estimation, and the result of the kernel density estimation is : ; Wherein, x is any candidate charging duration value point.

[0060] The result of the kernel density estimation is determined as the probability density function, and the charging duration feature is obtained.

[0061] Step S103, according to the charging start time feature and the charging duration feature, the behavior trajectory of the charging user is constructed; and based on the number feature and the behavior trajectory, Monte Carlo simulation is performed to construct multiple charging load scenarios.

[0062] In the embodiment, when the above Poisson distribution function, probability distribution model and probability density function are obtained, the behavior trajectory of the charging user can be constructed to describe the behavior rule of the charging user.

[0063] In some preferred embodiments, the behavior trajectory of the charging user is constructed according to the charging start time feature and the charging duration feature; and based on the number feature and the behavior trajectory, Monte Carlo simulation is performed to construct multiple charging load scenarios, including: according to the probability distribution model, the charging start time of the charging user is randomly extracted; according to the charging duration feature, the charging duration of the charging user is randomly extracted; according to the extracted charging start time and the extracted charging duration, the behavior trajectory of the charging user of each charging load scenario is constructed; based on the number feature, the number of charging users of each charging load scenario is randomly generated; based on the constructed behavior trajectory and the randomly generated number of charging users, Monte Carlo simulation is performed to construct multiple charging load scenarios.

[0064] Exemplarily, for each charging load scenario, the number of charging users, the charging start time and the charging duration therein can be randomly extracted from the Poisson distribution function, the probability distribution model and the probability density function in turn, respectively, and specifically represented as: The number of charging users of the s-th charging load scenario is randomly generated according to: ; For the s-th charging load scenario, the charging start time of the charging user is randomly extracted according to: . Wherein i refers to the i-th charging user, N sThe total number of charging users for the s-th charging load scenario.

[0065] The impulsive duration of the i-th charging user is d i : , and can be rounded to a time period.

[0066] Further, as shown in the following formula, the behavior trajectory of the charging user i in the charging load scenario s is constructed. Figure 9 : ; In the formula, when the charging user i is in the charging state in the charging load scenario s, the value is 1, and vice versa.

[0067] As shown in the following formula, the real-time charging power of the charging user i in the charging load scenario s is: Figure 10 ; In the formula, P is the rated power of the charging pile, and P is the real-time charging power of the charging user i in the charging load scenario s.

[0068] In some embodiments, the charging load scenario can be limited to meet the condition that the number of users charging at the same time cannot exceed the number of charging piles, which is represented as: ; Wherein, M is the number of charging piles.

[0069] At this time, all charging load scenarios containing complete information such as behavior trajectory, number of charging users, charging start time and charging duration are obtained.

[0070] Step S104, under the preset constraint condition, a preset solver is called to solve each charging load scenario respectively to generate a load prediction curve of each charging scenario; and the charging parameters of the charging station are regulated according to the load prediction curve.

[0071] In some embodiments, the preset constraint can include a charging pile power constraint, a charging service fee constraint and a user satisfaction constraint.

[0072] Wherein, the charging pile power constraint can refer to the above .

[0073] The charging service fee constraint can be represented as: . Wherein, is the charging service fee of the t-th time period, c min is the lower limit value of the preset service fee, and c max ​​​​​​is a preset upper limit value of the service fee.

[0074] The user satisfaction constraint can specifically be that the satisfaction degree of the charging user can not be lower than the existing pricing scheme, which can be expressed as, for example: ; Wherein: ; ; ; ; ; In the formula, represents the charging satisfaction degree of the charging user under the s-th charging load scenario, is the average satisfaction degree of the S charging load scenarios, represents the unit cost of charging of the charging user, including the real-time electricity price and the real-time charging service fee (which can be taken as an optimization variable), is the charging fee paid by the charging user for charging all charging demands at the lowest charging cost of the day under the s-th charging load scenario, is the charging fee paid by the charging user for charging all charging demands at the highest charging cost of the day under the s-th charging load scenario, is the average user satisfaction degree calculated based on the current pricing scheme and historical charging operation data, refers to the time variation amount.

[0075] Under the preset charging pile power constraint, charging service fee constraint and user satisfaction constraint, a commercial solver such as Cplex solver or Gurobi solver can be used to solve each charging load scenario respectively to generate the load prediction curve of each charging scenario.

[0076] During the solving, the objective function can be set as needed so that the solved load prediction curve can be more conducive to achieving certain goals. For example, for a charging operator, the goal can be to optimize the charging service fee of each time period to maximize the daily expected income under multiple scenarios: ; At this time, the load prediction curve obtained under this target constraint can be verified to obtain the optimal charging service fee strategy and the corresponding optimal user charging load curve, which is convenient for realizing the time-sharing dynamic and refined charging service fee design.

[0077] When the plurality of load prediction curves are obtained, the charging parameters of the charging station can be regulated as needed, such as the number of open charging piles or the open time period of the charging piles, so as to ensure efficient and stable operation of the transformer substation or the power system, avoid extreme situations such as overload, and reduce the risk of power consumption. In addition, the charging parameter to be regulated can also be the charging service fee. Through the adjustment of the charging service fee, on the one hand, the income of the operator can be optimized and improved, and on the other hand, through the adjustment of the charging service fee, the time of the charging user to the station and the number of people to the station for charging can be indirectly adjusted through the above-mentioned price driving, so as to maximize the efficiency and benefit.

[0078] Correspondingly, as shown in Figure 11 The application further provides a charging load regulation system 1100, which comprises a historical data acquisition module 1101, a feature extraction module 1102, a simulation scenario construction module 1103, and a regulation module 1104; wherein, The historical data acquisition module 1101 is configured to acquire historical operation data of a charging station; the historical operation data comprises a charging event sample of each charging user for each charging; The feature extraction module 1102 is configured to extract a number feature, a charging start time feature, and a charging duration feature of a charging user based on the charging event sample; The simulation scenario construction module 1103 is configured to construct a behavior trajectory of the charging user according to the charging start time feature and the charging duration feature; and perform Monte Carlo simulation based on the number feature and the behavior trajectory to construct a plurality of charging load scenarios; The regulation module 1104 is configured to call a preset solver to solve each charging load scenario under a preset constraint condition, to generate a load prediction curve of each charging scenario; and regulate charging parameters of the charging station according to the load prediction curves.

[0079] As a preferred solution, the charging station comprises a plurality of charging piles; the regulation system 1100 comprises a charging event generation module, which is configured to, before acquiring the historical operation data of the charging station: acquire running state data of the plurality of charging piles through a preset sensor; extract running records of each charging pile based on the running state data; generate a charging event each time a charging behavior of a charging user is monitored according to the running records, and construct the generated charging events into the charging event sample.

[0080] As a preferred solution, the feature extraction module 1102 extracts the number of charging users, the charging start time feature and the charging duration feature based on the charging event samples, including: The feature extraction module 1102 divides each charging event sample by time to obtain sample data of multiple discrete time periods; The user feature data is extracted from the sample data of the multiple discrete time periods by taking the preset data field as a retrieval key; wherein the user feature data is represented in a specific data type; The number of charging users, the charging start time feature and the charging duration feature are extracted from the user feature data.

[0081] As a preferred solution, the feature extraction module 1102 extracts the number of charging users, the charging start time feature and the charging duration feature from the user feature data, including: The feature extraction module 1102 constructs a Poisson distribution function of the number of charging users based on the user feature data, and determines the Poisson distribution function as the number feature; Based on the user feature data, a probability distribution model of the charging start time of the charging user is constructed, and the probability distribution model is determined as the charging start time feature; Based on the user feature data, a probability density function of the charging duration of the charging user is fitted, and the probability density function is determined as the charging duration feature.

[0082] As a preferred solution, the simulation scenario construction module 1103 constructs the behavior trajectory of the charging user according to the charging start time feature and the charging duration feature; and performs Monte Carlo simulation based on the number feature and the behavior trajectory to construct multiple charging load scenarios, including: The simulation scenario construction module 1103 randomly extracts the charging start time of the charging user according to the probability distribution model; and randomly extracts the charging duration of the charging user according to the charging duration feature; According to the extracted charging start time and the extracted charging duration, the behavior trajectory of the charging user of each charging load scenario is constructed; Based on the number feature, the number of charging users of each charging load scenario is randomly generated; Based on the constructed behavior trajectory and the randomly generated number of charging users, Monte Carlo simulation is performed to construct multiple charging load scenarios.

[0083] As a preferred solution, the feature extraction module 1102 constructs a probability distribution model of the charging start time of the charging user based on the user feature data, including: The feature extraction module 1102 calculates a first probability of the user starting charging based on the user feature data and a time dimension; A preset price sensitivity parameter is acquired; and a second probability of the user starting charging is calculated based on the user feature data, the price sensitivity parameter and a price dimension; A joint probability distribution is calculated based on the first probability and the second probability, and the probability distribution model is obtained.

[0084] Preferably, the user feature data includes a mean value of charging duration of sample data; and the feature extraction module 1102 fits a probability density function of charging duration of the charging user based on the user feature data, including: The feature extraction module 1102 calculates a sample standard deviation according to the mean value and each charging duration in the sample data; A bandwidth is calculated by a Silverma rule based on the standard deviation and a sample number; A Gaussian kernel function is obtained by kernel density estimation according to the sample number, the bandwidth and each charging duration in the sample data; The Gaussian kernel function is determined as the probability density function.

[0085] Preferably, the regulation module 1104 calls a preset solver to solve each charging load scenario under a preset constraint condition, and generates a load prediction curve of each charging scenario, including: The regulation module 1104 adopts a Cplex solver or a Gurobi solver to solve each charging load scenario under a preset charging pile power constraint, a charging service fee constraint and a user satisfaction constraint, and generates a load prediction curve of each charging scenario.

[0086] Correspondingly, the present application also provides a terminal device, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the regulation method of the charging load when executing the computer program.

[0087] The processor can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor. The processor is a control center of the terminal, and connects all parts of the terminal through various interfaces and lines.

[0088] The memory can be used to store the computer program, and the processor realizes various functions of the terminal by running or executing the computer program stored in the memory and calling data stored in the memory. The memory can mainly include a program storage area and a data storage area. The program storage area can store an operating system, at least one application program required for a function (such as a sound playing function, an image playing function, etc.), etc. The data storage area can store data created according to the use of the terminal (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a nonvolatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state memory devices.

[0089] Correspondingly, the present application further provides a computer readable storage medium, which comprises a stored computer program, wherein the computer program controls a device where the computer readable storage medium is located to execute the charging load regulation method when the computer program runs.

[0090] If the modules of the charging load regulation system are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above-mentioned embodiments can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer-readable storage medium. When the processor executes the computer program, the steps of each method embodiment described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or some intermediate forms, etc. The computer-readable medium can include any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium, etc.

[0091] Compared with the prior art, the present application has the following beneficial effects: The present application provides a charging load regulation method, system, terminal device and medium. By extracting features from charging event samples, the number of features, the charging start time feature and the electrical duration feature are obtained, and then a plurality of charging load scenarios are generated through Monte Carlo simulation. Compared with the prior art, the plurality of charging load scenarios can cover scenarios with random fluctuations in the number of users, charging start time, charging duration, etc. in actual operation, that is, considering the differences in user behavior leading to differences in charging time, the present application solves the one-sidedness of the prior art charging load prediction to some extent, effectively reduces the deviation between the charging load and the actual load scenario, and optimizes the authenticity of load prediction.

[0092] In addition, by calling a preset solver to solve each charging load scenario, a load prediction curve (i.e. multiple load prediction curves) of each charging load scenario is generated. Compared with the prior art technical solution of generating only a single curve (a single load curve is often based on the assumption of stable user behavior and only reflects the average law of load, and cannot cover extreme load or sudden fluctuations, etc.), the present application can cover load fluctuations under different user behavior trajectories and reflect the random fluctuation characteristics of load, making the prediction result further closer to the real scenario of charging operation.

[0093] Finally, regulating the charging parameters of the charging station according to each load prediction curve can ensure the efficient and stable operation of the charging station on the basis of accurate and real load prediction, and reduce the risk brought by extreme load or sudden large fluctuations, etc.

[0094] The above-described specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above-described specific embodiments are merely examples of the present application and are not intended to limit the protection scope of the present application. It is particularly pointed out that any modification, equivalent replacement, improvement, etc. made by those skilled in the art within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method of regulating a charging load, characterized by, The method comprises the following steps: acquiring historical operation data of a charging station; the historical operation data comprises charging event samples of each charging user for each charging; based on the charging event samples, extracting the number of charging users, charging start time characteristics and charging duration characteristics; constructing the behavior trajectory of the charging user according to the charging start time characteristics and the charging duration characteristics; and based on the number of characteristics and the behavior trajectory, performing Monte Carlo simulation to construct multiple charging load scenarios; under a preset constraint condition, calling a preset solver to solve each charging load scenario to generate a load prediction curve of each charging scenario; and according to the load prediction curve, adjusting and controlling the charging parameters of the charging station.

2. A method of regulating a charge load as claimed in claim 1, characterized in that, The charging station comprises multiple charging piles; Before acquiring the historical operation data of the charging station, the method comprises the following steps: collecting the running state data of the multiple charging piles through a preset sensor; based on the running state data, extracting the running record of each charging pile; according to the running record, generating a charging event each time the charging behavior of a charging user is monitored, and constructing the generated charging event as the charging event sample.

3. A method of regulating charge load as claimed in claim 2, wherein, The method comprises the following steps: dividing each charging event sample according to time to obtain sample data of multiple discrete time periods; extracting user feature data from the sample data of the multiple discrete time periods by taking a preset data field as a search key; wherein the user feature data is represented in a specific data type; extracting the number of charging users, charging start time characteristics and charging duration characteristics from the user feature data.

4. A method of regulating a charge load as claimed in claim 3, characterized in that, The method comprises the following steps: based on the user feature data, constructing a Poisson distribution function of the number of charging users, and determining the Poisson distribution function as the number of characteristics; based on the user feature data, constructing a probability distribution model of the charging start time of the charging user, and determining the probability distribution model as the charging start time characteristics; based on the user feature data, fitting a probability density function of the charging duration of the charging user, and determining the probability density function as the charging duration characteristics.

5. A method of regulating a charge load as claimed in claim 4, characterized in that, The method comprises the following steps: according to the probability distribution model, randomly extracting the charging start time of the charging user; according to the charging duration characteristics, randomly extracting the charging duration of the charging user; according to the extracted charging start time and the extracted charging duration, constructing the behavior trajectory of the charging user of each charging load scenario; based on the number of characteristics, randomly generating the number of charging users of each charging load scenario; based on the constructed behavior trajectory and the randomly generated number of charging users, performing Monte Carlo simulation to construct multiple charging load scenarios.

6. A method of regulating charge load as claimed in claim 4, wherein, The probability distribution model of the charging start time of the charging user is constructed based on the user feature data, including: Based on the user feature data and time dimension, the first probability of the user starting charging is calculated; A preset price sensitivity parameter is obtained; based on the user feature data, the price sensitivity parameter and the electricity price dimension, the second probability of the user starting charging is calculated; Based on the first probability and the second probability, a joint probability distribution is calculated to obtain the probability distribution model.

7. A method of regulating charge load as claimed in claim 4, wherein, The user feature data includes the mean of the charging duration of the sample data; the probability density function of the charging duration of the charging user is fitted based on the user feature data, including: According to the mean and each charging duration in the sample data, the sample standard deviation is calculated; Based on the standard deviation and the sample quantity, the bandwidth is calculated by the Silverma rule; According to the sample quantity, the bandwidth and each charging duration in the sample data, kernel density estimation is performed to obtain a Gaussian kernel function; The Gaussian kernel function is determined as the probability density function.

8. A method of regulating charge load according to any one of claims 1 to 7, characterized in that, Under the preset constraint condition, a preset solver is called to solve each charging load scenario to generate a load prediction curve of each charging scenario, including: Under the preset charging pile power constraint, charging service fee constraint and user satisfaction constraint, a Cplex solver or a Gurobi solver is used to solve each charging load scenario to generate a load prediction curve of each charging scenario.

9. A charging load regulation system, characterized in that, It includes a historical data acquisition module, a feature extraction module, a simulation scenario construction module and a regulation and control module; wherein, The historical data acquisition module is used to acquire the historical operation data of the charging station; the historical operation data includes the charging event sample of each charging user each time; The feature extraction module is used to extract the number feature, charging start time feature and charging duration feature of the charging user based on the charging event sample; The simulation scenario construction module is used to construct the behavior trajectory of the charging user according to the charging start time feature and the charging duration feature; and perform Monte Carlo simulation based on the number feature and the behavior trajectory to construct multiple charging load scenarios; The regulation and control module is used to call a preset solver to solve each charging load scenario under a preset constraint condition to generate a load prediction curve of each charging scenario; and regulate and control the charging parameters of the charging station according to each load prediction curve.

10. A terminal device, comprising: It includes a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, and the processor implements the charging load regulation and control method according to any one of claims 1 to 7 when executing the computer program.

Images