Charging station short-term load prediction method and system, electronic equipment and storage medium

Abstract

The invention provides a charging station short-term load prediction method and system, electronic equipment and a storage medium, and belongs to the technical field of electric traffic and power load prediction, and the method comprises the steps: obtaining historical charging load data, charging records and charging feature data of a target charging station, arranging according to time periods to form a historical charging load data set, a charging characteristic data set and charging automobile quantity data of each time period; constructing and training a charging automobile number prediction model, and predicting the number of charging automobiles; constructing and training a charging station load prediction model, predicting the load of the charging station, and forming a charging load prediction data set; and processing data in the charging load prediction data set into a charging load prediction curve. According to the method, various factors influencing the charging load are comprehensively considered, the actual charging condition can be accurately reflected, and the accuracy of short-term load prediction of the charging station is improved; the change trend of the predicted value of the charging load can be visually displayed, and the operability in practical application is enhanced.

Description

Background Art

[0002] With the popularity of electric vehicles, the number of charging piles is also increasing rapidly. The installation and use of a large number of charging piles has brought challenges to the safety and stability of the power grid. In order to improve power efficiency and reduce the overall power load, power supply units have the need to optimize resource allocation and balance the supply and demand of electricity. The formulation of relevant plans needs to rely on load estimation in the power consumption area. Therefore, accurate short-term power load forecasting of charging piles and appropriate evaluation of charging pile power consumption levels are crucial to building a safe and stable power grid.

[0003] At present, the prediction methods in the field of power load can be generally divided into long-term load prediction, medium-term load prediction and short-term load prediction according to the prediction time span. The time span of short-term load prediction is generally within one month, which is mainly used for real-time safety analysis and scheduling of power systems to meet power generation needs and ensure safe and stable operation of power systems. In the prior art, the short-term load prediction method of charging stations is usually based on the statistical analysis of the historical charging data of charging stations, such as collecting data such as the number of charging times, charging time and charging power of charging stations over a period of time, and then fitting through some common mathematical models, such as linear regression models, to obtain short-term charging station load prediction results.

[0004] However, in actual operation, the charging situation of charging piles is affected by many external factors, and its short-term load changes are unstable. The existing short-term load prediction methods for charging piles have a single data source and a relatively simple prediction model. It is difficult to fully capture the various complex factors affecting the charging load and their interrelationships, and it is impossible to accurately feedback the actual charging needs of electric vehicle users, resulting in low short-term prediction accuracy. At the same time, it is insufficient in processing complex nonlinear relationships, unable to accurately characterize the dynamic change trend of the charging load, and unable to provide a reliable basis in terms of resource optimization and allocation, coping with peak loads, etc., and it is difficult to meet the demand for accurate prediction in charging pile operation and management. Summary of the invention

[0005] In view of the technical problems that the existing short-term load forecasting methods for charging stations have a single data source, a simple model, low forecasting accuracy, difficulty in handling complex nonlinear relationships, inability to accurately characterize changing trends and meet the precise forecasting needs of operation management, the present invention provides a short-term load forecasting method, system, electronic device and storage medium for charging stations, which comprehensively consider various factors affecting the charging load and sequentially construct a charging car quantity forecasting model and a charging station load forecasting model for forecasting, so that the forecasting results can accurately reflect the actual charging situation and improve the accuracy of the short-term load forecasting of the charging station; the charging load forecasting data is processed into a smooth curve to intuitively display the changing trend, thereby enhancing the operability in practical applications.

[0006] In a first aspect, the present invention provides a method for short-term load forecasting of a charging station, the steps comprising: S1. Obtain the historical charging load data, charging records and charging characteristic data corresponding to the charging records of the target charging station, the charging characteristic data includes charging user characteristic information, charging vehicle characteristic information, charging station characteristic information, traffic flow data around the charging station, time-of-use electricity price information, and charging time data; S2. Arrange the historical charging load data, charging records and charging characteristic data by time period to obtain the historical charging load data set, charging characteristic data set and charging vehicle quantity data for each time period of the target charging station; The steps for obtaining the data on the number of charging vehicles are as follows: S201. For each charging record, the entire period between the charging start time and the charging end time is defined as the charging period. During the charging period of the charging record, the corresponding electric vehicle is in a charging state; S202. Count the number of electric vehicles in the charging state at the target charging station in each period to form the number of charging vehicles in that period; S3. Obtaining the external environment characteristic data around the target charging station and sorting it by time period to form an external environment characteristic data set for each time period, the external environment characteristic data including weather information and temperature information; S4. Using the charging characteristic dataset and the external environment characteristic dataset of each time period as the variable set and the charging vehicle quantity data of each time period as the result set to construct and train a charging vehicle quantity prediction model, the output value of the charging vehicle quantity prediction model is the predicted value of the charging vehicle quantity of the target charging station in each time period; S5. The charging characteristic data set for each time period and the predicted value of the external environment characteristic data for each time period within a period of time are input into the charging vehicle quantity prediction model to obtain the predicted value of the number of charging vehicles at the target charging station for each time period within a period of time to form a charging vehicle quantity prediction data set; S6. Use the number of charging vehicles in each period, the characteristic information of the charging station, and the historical charging load data set as a training set to construct and train a charging station load prediction model, wherein the output value of the charging station load prediction model is the predicted value of the charging load in the period to be predicted; S7. Input the charging vehicle quantity prediction data set and the charging station characteristic information into the charging station load prediction model to obtain the charging load prediction data of the target charging station for each period of time to form a charging load prediction data set; S8. Process the data in the charging load prediction data set into a charging load prediction curve, where the charging load prediction curve is a smooth curve showing the charging load data changing over time.

[0007] It should be further noted that, in step S1, the charging record includes the charging start time and the charging end time; Charging user characteristic information includes charging record ID, user's monthly charging amount, and user's charging discount; Charging vehicle characteristic information includes vehicle type and vehicle battery capacity; Charging station characteristic information includes the location of the charging station, the type of charging station, the maximum power of the charging station, the total number of charging piles in the charging station, the number of DC charging piles, and the number of AC charging piles; The charging time data includes the charging date and whether the charging date is a holiday.

[0008] It should be further explained that, in step S2, a day is divided into 96 time periods, and the duration of each time period is 15 minutes.

[0009] It should be further explained that, in step S2, the historical charging load data, charging records and charging characteristic data are first standardized and then sorted by time period; In step S3, the external environment characteristic data around the target charging station is first standardized and then sorted by time period; The specific operation of the standardization process is: encoding the categorical feature data, and then normalizing the encoded values. The encoding method includes one-hot encoding; directly normalizing the numerical feature data.

[0010] It should be further explained that the encoded numerical value or numerical feature data is normalized using the Min-Max normalization equation, and the expression of the Min-Max normalization equation is:

[0011] In the formula, It is the data after normalization, which is a value between 0 and 1; is the original data; is the minimum value in the feature data set to which the original data belongs; It is the maximum value in the feature data set to which the original data belongs.

[0012] It should be further explained that in step S4, the XGBoost model is used to build and train the charging car quantity prediction model, and the additive model is used to sequentially build and superimpose multiple decision trees to fit the residuals.

[0013] It should be further explained that, in step S5, the LSTM model is used to construct and train the charging station load prediction model.

[0014] Further preferably, step S5 inputs the predicted values ​​of the external environment characteristic data in each time period in the next 24 hours into the charging vehicle quantity prediction model, obtains the predicted values ​​of the charging vehicle quantity in each time period in the next 24 hours at the target charging station, and forms a charging vehicle quantity prediction data set; Step S7 obtains charging load forecast data for each time period in the next 24 hours at the target charging station.

[0015] It should be further explained that, in step S8, a Gaussian smoothing algorithm is used to process the data in the charging load prediction data set.

[0016] In a second aspect, the present invention provides a charging station short-term load forecasting system, which is used to implement the above-mentioned charging station short-term load forecasting method, including: A data acquisition module is used to acquire historical charging load data, charging records, charging characteristic data corresponding to the charging records, and external environment characteristic data around the charging station of the target charging station; The data preprocessing module is used to sort the historical charging load data, charging records and charging feature data by time period to form a historical charging load data set, a charging feature data set and charging vehicle quantity data for each time period of the target charging station; sort the external environment feature data around the target charging station by time period to form an external environment feature data set for each time period; A charging vehicle quantity prediction model construction and training module is used to construct and train a charging vehicle quantity prediction model; A charging vehicle quantity prediction module is used to obtain a charging vehicle quantity prediction value using a charging vehicle quantity prediction model to form a charging vehicle quantity prediction data set; The charging station load prediction model construction and training module is used to construct and train the charging station load prediction model; A charging station load prediction module is used to use a charging station load prediction model to predict charging load prediction data of a target charging station to form a charging load prediction data set; A data smoothing processing module is used to process the data in the charging load prediction data set into a smooth curve showing the charging load data changing over time; The result display module is used to display the charging load prediction curve and corresponding data in the form of graphics and data.

[0017] In a third aspect, the present invention provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is configured to implement the steps of the above-mentioned charging station short-term load forecasting method when executing the computer program.

[0018] In a fourth aspect, the present invention provides a storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the above-mentioned charging station short-term load forecasting method are implemented.

[0019] The beneficial effects of the present invention are: 1. The charging station short-term load prediction method of the present invention integrates the historical charging load data charging records, charging characteristic data and external environment characteristic data and organizes them by time period, sequentially constructs charging vehicle quantity prediction models and charging station load prediction models in different time periods, predicts the number of charging vehicles and charging load within a period of time, and finally obtains the charging load prediction curve. The present invention comprehensively considers various factors affecting the charging load, avoids the limitations of using one or a few factors for prediction, and enables the prediction results to more accurately reflect the actual charging situation of the charging station; dividing the data by time period helps to capture the changing rules of the charging load in a short period of time, further improves the prediction accuracy, and can improve the accuracy of the short-term load prediction of the charging station.

[0020] 2. The present invention ultimately processes the charging load prediction data into a smooth curve, and displays the changing trend of the charging load over time in an intuitive graphical manner, which makes it easier for people to understand the prediction results. Thus, according to the peak and trough periods of the curve, the use of charging piles can be reasonably arranged, resources can be allocated, and peak-shifting charging strategies can be formulated, thereby enhancing the operability of the entire prediction method in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings required for use in the description will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0022] Figure 1 It is a flow chart of a method for short-term load prediction of a charging station in one embodiment of the present invention.

[0023] Figure 2 It is a schematic block diagram of a short-term load forecasting system for a charging station in one embodiment of the present invention.

[0024] Figure 3 It is a schematic diagram of the hardware structure of an electronic device in one embodiment of the present invention. DETAILED DESCRIPTION

[0025] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0026] The short-term load forecasting method of the charging station involved in this application is mainly aimed at the technical field of electric transportation and power load forecasting. The beneficial effect is that the charging records of historical charging load data, charging characteristic data and external environment characteristic data are integrated and sorted by time period, and the charging vehicle quantity forecasting model and charging station load forecasting model of different time periods are constructed in turn, and the charging vehicle quantity and charging load within a period of time are predicted, and finally the charging load forecasting curve is obtained. The present invention comprehensively considers various factors affecting the charging load, avoids the limitation of using one or a few factors for prediction, so that the prediction result can more accurately reflect the actual charging situation of the charging station; dividing the data by time period helps to capture the changing law of the charging load in a short time, further improves the prediction accuracy, and can improve the accuracy of the short-term load forecasting of the charging station; processing the charging load prediction data into a smooth curve, and displaying the changing trend of the charging load over time in an intuitive graphical way, can make it easier for people to understand the prediction results, so that according to the peak and trough periods of the curve, the use of charging piles can be reasonably arranged, resources can be allocated, and peak-shifting charging strategies can be formulated, etc., which enhances the operability of the entire prediction method in practical applications.

[0027] The short-term load forecasting method for charging stations involved in this application mainly targets the technical problems of existing short-term load forecasting methods for charging stations, such as single data source, simple model, low prediction accuracy, difficulty in handling complex nonlinear relationships, inability to accurately characterize changing trends and meet the precise prediction needs of operation management.

[0028] The following is a detailed description of the charging station short-term load forecasting method involved in the present application. For the purpose of illustration rather than limitation, specific details such as specific system structures and technologies are proposed to facilitate a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application can also be implemented in other embodiments without these specific details.

[0029] In the charging station short-term load forecasting method involved in the present application, the term "include" used indicates the existence of the described features, entities, steps, operations, elements and / or components, but does not exclude the existence or addition of one or more other features, entities, steps, operations, elements, components and / or their collections. The terms "include", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0030] In order to clearly describe the technical solution of the present application, the words "first", "second" and the like are used to distinguish the same or similar items with substantially the same functions and effects. Those skilled in the art can understand that the words "first", "second" and the like do not limit the quantity and execution order, and the words "first", "second" and the like do not necessarily limit the difference.

[0031] The phrases such as "one embodiment" or "some embodiments" described in the present application mean that the specific features, structures or characteristics described in the embodiment are included in one or more embodiments of the present application. Therefore, the phrases such as "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments" etc. that appear in different places in the present application do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways.

[0032] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] The charging station short-term load forecasting method provided by the embodiment of the present invention is executed by a computer device, and accordingly, the charging station short-term load forecasting system runs in the computer device.

[0034] Figure 1 : is a flow chart of a method for short-term load forecasting of a charging station according to an embodiment of the present invention. Figure 1 The execution subject may be a short-term load forecasting system for a charging station. According to different requirements, the order of the steps in the flow chart may be changed, and some may be omitted.

[0035] like Figure 1 As shown, the short-term load forecasting method for the charging station includes: Step S1. Obtain historical charging load data, charging records, and charging characteristic data corresponding to the charging records of the target charging station, wherein the charging characteristic data includes charging user characteristic information, charging vehicle characteristic information, charging station characteristic information, traffic flow data around the charging station, time-of-use electricity price information, and charging time data.

[0036] By acquiring a variety of data, including historical charging load data, charging records, and charging characteristic data, a comprehensive data foundation is provided for subsequent predictions. These data can reflect the impact of many factors on charging load, such as charging station usage habits, user characteristics, vehicle conditions, surrounding traffic, and electricity prices, thereby improving the accuracy of the prediction.

[0037] In some embodiments, the charging record includes a charging start time and a charging end time; Charging user characteristic information includes charging record ID, user's monthly charging amount, and user's charging discount; Charging vehicle characteristic information includes vehicle type and vehicle battery capacity; Charging station characteristic information includes the location of the charging station, the type of charging station, the maximum power of the charging station, the total number of charging piles in the charging station, the number of DC charging piles, and the number of AC charging piles; The charging time data includes the charging date and whether the charging date is a holiday.

[0038] By defining the specific contents of the charging record and various characteristic information in detail, step S1 is made more specific and operable.

[0039] Step S2, sorting the historical charging load data, charging records and charging characteristic data by time period to obtain the historical charging load data set, charging characteristic data set and charging vehicle quantity data of each time period of the target charging station; The steps for obtaining the data on the number of charging vehicles are as follows: S201. For each charging record, the entire period between the charging start time and the charging end time is defined as the charging period. During the charging period of the charging record, the corresponding electric vehicle is in a charging state; S202. Count the number of electric vehicles in a charging state at the target charging station in each time period to form data on the number of charging vehicles in the time period.

[0040] Organizing data by time period to form different data sets is conducive to analyzing the charging patterns in different time periods, and then carefully capturing the changing trend of charging load within a day, so that subsequent models can better learn the charging behavior characteristics at different time intervals.

[0041] In some embodiments, a day is divided into 96 equal periods, and the duration of each period is 15 minutes.

[0042] The 15-minute time period division can more finely analyze the time distribution of charging behavior. Compared with the coarse time period division, it can better capture the fluctuation of charging load and help improve the accuracy of prediction.

[0043] In some embodiments, the historical charging load data, charging records, and charging characteristic data are first standardized and then sorted by time period; The specific operation of the standardization process is: encoding the categorical feature data, and then normalizing the encoded values. The encoding method includes one-hot encoding; directly normalizing the numerical feature data.

[0044] In some embodiments, the encoded numerical value or numerical feature data is normalized using a Min-Max normalization equation, and the expression of the Min-Max normalization equation is:

[0045] In the formula, It is the data after normalization, which is a value between 0 and 1; is the original data; is the minimum value in the feature data set to which the original data belongs; It is the maximum value in the feature data set to which the original data belongs.

[0046] Processing categorical and numerical data separately can eliminate the dimensional differences between data, make the data on the same scale, and avoid some features having too much impact on model training due to too large a numerical range, thereby improving the efficiency and accuracy of model training.

[0047] Step S3, obtaining external environment characteristic data around the target charging station and sorting them by time period to form an external environment characteristic data set for each time period, wherein the external environment characteristic data includes weather information and temperature information.

[0048] Acquire and organize external environmental characteristic data, taking into account the impact of factors such as weather and temperature on charging demand. For example, in hot or cold weather, the battery performance of electric vehicles may be affected, thus changing the user's charging habits. Therefore, fully considering the external environmental characteristic data can make the prediction model more in line with the actual situation.

[0049] In some embodiments, the external environment characteristic data around the target charging station is first standardized and then sorted by time period.

[0050] Step S4, using the charging characteristic data set and the external environment characteristic data set of each time period as a variable set and the charging vehicle quantity data of each time period as a result set to construct and train a charging vehicle quantity prediction model, the output value of the charging vehicle quantity prediction model is the predicted value of the charging vehicle quantity at the target charging station in each time period.

[0051] Construct and train a charging vehicle number prediction model, with charging characteristics and external environment characteristics as variables and the number of charging vehicles as the result. This model can explore the potential relationship between these factors and the number of charging vehicles, and can predict the number of charging vehicles in each time period based on different conditions, providing important intermediate data for subsequent charging load prediction.

[0052] In some embodiments, an XGBoost model is used to build and train a charging car quantity prediction model, and an additive model is used to sequentially build and overlay multiple decision trees to fit the residuals.

[0053] The XGBoost model is efficient and accurate. By sequentially constructing and stacking multiple decision trees to fit the residuals, it can handle complex nonlinear relationships well and effectively mine the complex relationship between the number of charging cars and various features, thereby improving the accuracy of predictions.

[0054] In some embodiments, the objective function of the XGBoost model is expressed as:

[0055] In the formula, is the loss function, usually squared error ; is a regular term used to control model complexity and prevent overfitting; k is the number of decision trees, for each decision tree , Zheng Zexiang is usually the sum of the number of leaf nodes and the square of the leaf node weight; The gain formula of the XGBoost model is expressed as:

[0056] In the formula, is the gradient, that is, the first-order derivative of the residual, indicating the error direction and magnitude of each sample; is the sum of the second-order derivatives of the residual, which is used to adjust the size of the gain; and is a regularization hyperparameter.

[0057] Step S5, input the charging characteristic data set of each time period and the predicted value of the external environment characteristic data of each time period within a period of time into the charging vehicle quantity prediction model, obtain the predicted value of the charging vehicle quantity of each time period within a period of time at the target charging station, and form a charging vehicle quantity prediction data set.

[0058] By using the trained charging vehicle number prediction model combined with the predicted values ​​of future external environment characteristics, the predicted values ​​of the number of charging vehicles in each period within a period of time can be obtained. The number of charging vehicles in each period can be estimated in advance, providing data support for the operation and management of charging stations.

[0059] In some embodiments, step S5 is to input the predicted values ​​of the external environment characteristic data for each time period in the next 24 hours into the charging vehicle quantity prediction model, obtain the predicted values ​​of the charging vehicle quantity for each time period in the next 24 hours at the target charging station, and form a charging vehicle quantity prediction data set.

[0060] The accuracy of the prediction value of the number of charging vehicles is strongly correlated with the accuracy of the prediction value of the external environment characteristic data input into the charging vehicle number prediction model. In the prior art, the accuracy of predicting the external environment characteristic data within the next 24 hours is relatively high, so the accuracy of inputting it into the charging number prediction model for prediction is also relatively high.

[0061] Step S6, using the charging car quantity data, charging station feature information and historical charging load data set in each time period as a training set to build and train a charging station load prediction model, the output value of the charging station load prediction model is the charging load prediction value of the time period to be predicted.

[0062] Constructing and training a charging station load prediction model, using the number of charging vehicles in each time period, charging station characteristic information and historical charging load data set as the training set, can comprehensively consider multiple factors to predict the charging load, effectively utilize previously collected and organized data, and enable the model to more accurately predict the charging load in each time period.

[0063] In some embodiments, a charging station load prediction model is constructed and trained using an LSTM model.

[0064] The LSTM model has advantages in processing time series data. Since the charging load data changes over time, LSTM can well capture the long-term dependencies in the time series, thereby more accurately predicting future charging loads.

[0065] In some embodiments, the steps of constructing the charging station load prediction model are: S601: The charging feature data set, the charging station feature information, and the historical charging load data set are sorted by time period and standardized, and then used as input data of the LSTM network; S602. Setting one or more layers of LSTM network to learn the temporal dependency in the input data, each LSTM layer will capture the short-term dependency and long-term dependency of the input data; S603. Set a fully connected layer to output the charging load prediction value for the time period to be predicted.

[0066] In some embodiments, historical data is used to perform a training step on the charging station load prediction model, batch data is used during training to update the parameters of the charging station load prediction model, the optimization algorithm used in the training process is the Adam optimizer, and the loss function is the mean square error.

[0067] Step S7, inputting the charging vehicle quantity prediction data set and the charging station characteristic information into the charging station load prediction model, obtaining the charging load prediction data of each time period of the target charging station within a period of time, and forming a charging load prediction data set.

[0068] The charging load forecast data output by the charging station load forecast model can accurately predict the charging load forecast data for each period in the future, which helps to reasonably arrange charging pile maintenance, optimize power resource allocation, respond to peak demand in advance, effectively improve operational efficiency and service quality, and enhance user charging experience.

[0069] In some embodiments, step S7 obtains charging load forecast data for each time period in the next 24 hours at the target charging station.

[0070] Step S8, processing the data in the charging load prediction data set into a charging load prediction curve, where the charging load prediction curve is a smooth curve showing the charging load data changing over time.

[0071] Processing the predicted charging load data into a smooth curve can intuitively display the changing trend of the charging load over time, allowing users to more clearly understand the dynamic changes of the charging load and conveniently observe the peak and trough periods of the load, providing a more intuitive basis for reasonable resource allocation and scheduling.

[0072] In some embodiments, the data within the charging load prediction dataset is processed using a Gaussian smoothing algorithm.

[0073] Using the Gaussian smoothing algorithm to process the data in the charging load prediction data set can reduce the noise in the data, making the generated load prediction curve smoother and more in line with the actual change trend. This can more intuitively display the overall change pattern of the charging load, which is convenient for users to observe and analyze.

[0074] In a specific embodiment, the charging station short-term load forecasting method includes: Step S1. Obtain the historical charging load data, charging records and charging characteristic data corresponding to the charging records of the target charging station, the charging characteristic data including charging user characteristic information, charging vehicle characteristic information, charging station characteristic information, traffic flow data around the charging station, time-of-use electricity price information, and charging time data; Among them, the charging record includes the charging start time and the charging end time; Charging user characteristic information includes charging record ID, user's monthly charging amount, and user's charging discount; Charging vehicle characteristic information includes vehicle type and vehicle battery capacity; Charging station characteristic information includes the location of the charging station, the type of charging station, the maximum power of the charging station, the total number of charging piles in the charging station, the number of DC charging piles, and the number of AC charging piles; Charging time data includes charging date and whether the charging date is a holiday; Step S2. Divide a day into 96 time periods, each of which lasts 15 minutes; The historical charging load data, charging records and charging characteristic data are standardized and sorted by time period to form the historical charging load data set, charging characteristic data set and charging vehicle quantity data for each time period of the target charging station; The steps for obtaining the data on the number of charging vehicles are as follows: S201. For each charging record, the entire period between the charging start time and the charging end time is defined as the charging period. During the charging period of the charging record, the corresponding electric vehicle is in a charging state; S202. Count the number of electric vehicles in the charging state at the target charging station in each period to form the number of charging vehicles in that period; The specific operations of the standardization process are: encoding the categorical feature data, and then normalizing the encoded values. The encoding method includes one-hot encoding; directly normalizing the numerical feature data; The Min-Max normalization equation is used to normalize the encoded numerical value or numerical feature data. The expression of the Min-Max normalization equation is:

[0075] In the formula, It is the data after normalization, which is a value between 0 and 1; is the original data; is the minimum value in the feature data set to which the original data belongs; is the maximum value in the feature data set to which the original data belongs; Step S3. Obtain the external environment characteristic data around the target charging station, standardize it and organize it by time period to form an external environment characteristic data set for each time period. The external environment characteristic data includes weather information and temperature information. The specific operation of the standardization process is the same as that in step S2. Step S4. The charging feature dataset and the external environment feature dataset of each time period are used as the variable set, and the data of the number of charging vehicles in each time period is used as the result set. The XGBoost model is used to build and train a charging vehicle number prediction model. Multiple decision trees are sequentially constructed and superimposed to fit the residual using an additive model. The output value of the charging vehicle number prediction model is the predicted value of the number of charging vehicles at the target charging station in each time period. The objective function of the XGBoost model is expressed as:

[0076] In the formula, is the loss function, usually squared error ; is a regular term used to control model complexity and prevent overfitting; k is the number of decision trees, for each decision tree , Zheng Zexiang is usually the sum of the number of leaf nodes and the square of the leaf node weight; The gain formula of the XGBoost model is expressed as:

[0077] In the formula, is the gradient, that is, the first-order derivative of the residual, indicating the error direction and magnitude of each sample; is the sum of the second-order derivatives of the residual, which is used to adjust the size of the gain; and is the regularization hyperparameter; Step S5. Input the charging characteristic data set of each time period and the predicted value of the external environment characteristic data of each time period in the next 24 hours into the charging vehicle quantity prediction model, obtain the predicted value of the number of charging vehicles in each time period in the next 24 hours at the target charging station, and form a charging vehicle quantity prediction data set; Step S6. Using the number of charging vehicles in each time period, the characteristic information of the charging station and the historical charging load data set as the training set, the LSTM model is used to build and train the charging station load prediction model, and the output value of the charging station load prediction model is the charging load prediction value of the time period to be predicted; The steps to construct the charging station load forecasting model are: S601: The charging feature data set, the charging station feature information, and the historical charging load data set are sorted by time period and standardized, and then used as input data of the LSTM network; S602. Setting one or more layers of LSTM network to learn the temporal dependency in the input data, each LSTM layer will capture the short-term dependency and long-term dependency of the input data; S603. Setting a fully connected layer to output the charging load prediction value for the predicted period; The charging station load prediction model is trained using historical data. During training, batch data is used to update the parameters of the charging station load prediction model. The optimization algorithm used in the training process is the Adam optimizer, and the loss function is the mean square error. Step S7. Input the charging vehicle quantity prediction data set and the charging station characteristic information into the charging station load prediction model to obtain the charging load prediction data of the target charging station for each period in the next 24 hours to form a charging load prediction data set; Step S8. The Gaussian smoothing algorithm processes the data in the charging load prediction data set into a charging load prediction curve. The charging load prediction curve is a smooth curve showing the charging load data changing over time.

[0078] The following is an embodiment of a short-term load forecasting system for a charging station provided in an embodiment of the present disclosure. The short-term load forecasting system for a charging station and the short-term load forecasting methods for a charging station in the above-mentioned embodiments belong to the same inventive concept. For details not described in detail in the embodiment of the short-term load forecasting system for a charging station, reference can be made to the embodiment of the short-term load forecasting method for a charging station in the above-mentioned embodiments.

[0079] A mobile terminal implementing various embodiments of the present invention will now be described with reference to the accompanying drawings. In the subsequent description, suffixes such as "module", "component" or "unit" used to represent elements are used only to facilitate the description of the embodiments of the present invention and have no specific meaning in themselves. Therefore, "module" and "component" can be used interchangeably.

[0080] like Figure 2 As shown, the short-term load forecasting system for charging stations includes: A data acquisition module is used to acquire historical charging load data, charging records, charging characteristic data corresponding to the charging records, and external environment characteristic data around the charging station of the target charging station; The data preprocessing module is used to sort the historical charging load data, charging records and charging feature data by time period to form a historical charging load data set, a charging feature data set and charging vehicle quantity data for each time period of the target charging station; sort the external environment feature data around the target charging station by time period to form an external environment feature data set for each time period; A charging vehicle quantity prediction model construction and training module is used to construct and train a charging vehicle quantity prediction model; A charging vehicle quantity prediction module is used to obtain a charging vehicle quantity prediction value using a charging vehicle quantity prediction model to form a charging vehicle quantity prediction data set; The charging station load prediction model construction and training module is used to construct and train the charging station load prediction model; A charging station load prediction module is used to use a charging station load prediction model to predict charging load prediction data of a target charging station to form a charging load prediction data set; A data smoothing processing module is used to process the data in the charging load prediction data set into a charging load prediction curve; The result display module is used to display the charging load prediction curve and corresponding data in the form of graphics and data.

[0081] Those skilled in the art will appreciate that the electronic device structure involved in the embodiments of the present invention does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0082] Figure 3 A schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0083] The electronic device 500 includes, but is not limited to, components such as a processor 501, a network module 502, an audio output unit 503, an input unit 504, a display unit 506, a user input unit 507, an interface unit 508, and a memory 509. Those skilled in the art will appreciate that the electronic device structure involved in the embodiments of the present invention does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0084] In the embodiments of the present invention, electronic devices include but are not limited to laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices can also represent various forms of mobile devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are only examples and are not intended to limit the implementation of the embodiments of the present application described and / or required herein.

[0085] In the embodiment of the present application, the processor 501 can be implemented by using at least one of an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a processor, a controller, a microcontroller, a microprocessor, and an electronic unit designed to perform the functions described herein. In some cases, such an implementation can be implemented in a controller. For software implementation, implementations such as processes or functions can be implemented with separate software modules that allow execution of at least one function or operation. The software code can be implemented by a software application (or program) written in any appropriate programming language, and the software code can be stored in a memory and executed by a controller.

[0086] The display unit 506 is used to display information input by the user or information provided to the user. The display unit 506 may include a display panel, which may be configured in the form of a liquid crystal display (LCD), an organic light-emitting diode (OLED), etc.

[0087] The user input unit 507 may include, but is not limited to, a physical keyboard, function keys (such as a volume control key, a switch key, etc.), a trackball, a mouse, and a joystick, which will not be described in detail here.

[0088] The interface unit 508 is an interface for connecting external devices to the electronic device 500. For example, the external devices may include a wired or wireless headset port, an external power supply (or battery charger) port, a wired or wireless data port, a memory card port, a port for connecting a device with an identification module, an audio input / output (I / O) port, a video I / O port, an earphone port, and the like.

[0089] In addition, the electronic device 500 includes some functional modules not shown, which will not be described in detail here.

[0090] Those skilled in the art will appreciate that the various aspects of the electronic device provided by the present application may be implemented as a system, method or program product. Therefore, the various aspects of the present disclosure may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software, which may be collectively referred to as a "circuit", "module" or "system" herein.

[0091] The present application also provides a storage medium, in which a program product capable of implementing a method for short-term load prediction of a charging station is stored. In some possible implementations, various aspects of the present disclosure may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary implementations of the present disclosure described in the above “Exemplary Method” section of this specification.

[0092] The storage medium can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0093] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for short-term load forecasting of a charging station, characterized in that the steps include: S1. Obtain the historical charging load data, charging records and charging characteristic data corresponding to the charging records of the target charging station, the charging characteristic data includes charging user characteristic information, charging vehicle characteristic information, charging station characteristic information, traffic flow data around the charging station, time-of-use electricity price information, and charging time data; S2. Arrange the historical charging load data, charging records and charging characteristic data by time period to obtain the historical charging load data set, charging characteristic data set and charging vehicle quantity data for each time period of the target charging station; The steps for obtaining the data on the number of charging vehicles are as follows: S201. For each charging record, the entire period between the charging start time and the charging end time is defined as the charging period. During the charging period of the charging record, the corresponding electric vehicle is in a charging state; S202. Count the number of electric vehicles in the charging state at the target charging station in each period to form the number of charging vehicles in that period; S3. Obtaining the external environment characteristic data around the target charging station and sorting it by time period to form an external environment characteristic data set for each time period, the external environment characteristic data including weather information and temperature information; S4. Using the charging characteristic dataset and the external environment characteristic dataset of each time period as the variable set and the charging vehicle quantity data of each time period as the result set to construct and train a charging vehicle quantity prediction model, the output value of the charging vehicle quantity prediction model is the predicted value of the charging vehicle quantity of the target charging station in each time period; S5. The charging characteristic data set for each time period and the predicted value of the external environment characteristic data for each time period within a period of time are input into the charging vehicle quantity prediction model to obtain the predicted value of the number of charging vehicles at the target charging station for each time period within a period of time to form a charging vehicle quantity prediction data set; S6. Use the number of charging vehicles in each period, the characteristic information of the charging station, and the historical charging load data set as a training set to construct and train a charging station load prediction model, wherein the output value of the charging station load prediction model is the predicted value of the charging load in the period to be predicted; S7. Input the charging vehicle quantity prediction data set and the charging station characteristic information into the charging station load prediction model to obtain the charging load prediction data of the target charging station for each period of time to form a charging load prediction data set; S8. Process the data in the charging load prediction data set into a charging load prediction curve, where the charging load prediction curve is a smooth curve showing the charging load data changing over time.

2. The method for short-term load forecasting of a charging station according to claim 1, characterized in that: In step S1, the charging record includes the charging start time and the charging end time; Charging user characteristic information includes charging record ID, user's monthly charging amount, and user's charging discount; Charging vehicle characteristic information includes vehicle type and vehicle battery capacity; Charging station characteristic information includes the location of the charging station, the type of charging station, the maximum power of the charging station, the total number of charging piles in the charging station, the number of DC charging piles, and the number of AC charging piles; The charging time data includes the charging date and whether the charging date is a holiday.

3. The method for short-term load forecasting of a charging station according to claim 1, characterized in that: In step S2, a day is divided into 96 time periods, and the duration of each time period is 15 minutes.

4. The method for short-term load forecasting of a charging station according to claim 1, characterized in that: In step S2, the historical charging load data, charging records and charging characteristic data are first standardized and then sorted by time period; In step S3, the external environment characteristic data around the target charging station is first standardized and then sorted by time period; The specific operation of the standardization process is: encoding the categorical feature data, and then normalizing the encoded values. The encoding method includes one-hot encoding; directly normalizing the numerical feature data.

5. The method for short-term load forecasting of a charging station according to claim 1, characterized in that: In step S4, the XGBoost model is used to build and train a charging car quantity prediction model, and multiple decision trees are sequentially built and superimposed using an additive model to fit the residuals.

6. The method for short-term load forecasting of a charging station according to claim 1, characterized in that: In step S5, the LSTM model is used to build and train the charging station load prediction model.

7. The method for short-term load forecasting of a charging station according to claim 1, characterized in that: In step S8, a Gaussian smoothing algorithm is used to process the data in the charging load prediction data set.

8. A short-term load forecasting system for a charging station, characterized in that: The method for short-term load prediction of a charging station according to any one of claims 1 to 7 comprises: A data acquisition module is used to acquire historical charging load data, charging records, charging characteristic data corresponding to the charging records, and external environment characteristic data around the charging station of the target charging station; The data preprocessing module is used to sort the historical charging load data, charging records and charging feature data by time period to form a historical charging load data set, a charging feature data set and charging vehicle quantity data for each time period of the target charging station; sort the external environment feature data around the target charging station by time period to form an external environment feature data set for each time period; A charging vehicle quantity prediction model construction and training module is used to construct and train a charging vehicle quantity prediction model; A charging vehicle quantity prediction module is used to obtain a charging vehicle quantity prediction value using a charging vehicle quantity prediction model to form a charging vehicle quantity prediction data set; The charging station load prediction model construction and training module is used to construct and train the charging station load prediction model; A charging station load prediction module is used to use a charging station load prediction model to predict charging load prediction data of a target charging station to form a charging load prediction data set; A data smoothing processing module is used to process the data in the charging load prediction data set into a charging load prediction curve; The result display module is used to display the charging load prediction curve and corresponding data in the form of graphics and data.

9. An electronic device, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor is used to implement the steps of the method for short-term load forecasting of a charging station as described in any one of claims 1 to 7 when executing the computer program.

10. A storage medium, characterized in that: A computer program is stored on the storage medium, and when the computer program is executed by the processor, the steps of the charging station short-term load forecasting method as described in any one of claims 1 to 7 are implemented.

Images