Ordered access scheduling method and device for charging load

Abstract

The invention provides an orderly access scheduling method and device for charging loads, and belongs to the technical field of electric power Internet of Things. The method comprises the following steps: inputting real-time time sequence data into a load prediction model to obtain load prediction time sequence data of a power distribution area output by the load prediction model; receiving an attribute parameter of a charging load node and a charging service request parameter; calculating the access probability of the charging load node based on the load prediction time sequence data and the attribute parameters; determining the time of the minimum load prediction time sequence data corresponding to the access probability of the maximum charging load node as the charging starting time, and determining the reserved charging load of the charging load node based on the charging starting time and the charging service request parameter; and calculating updated load prediction time sequence data of the power distribution area based on the reserved charging load and the load prediction time sequence data. The method is used for solving the problems of how to fully consider the power utilization side load and how to reduce the fluctuation influence of the charging load on the power distribution network load.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power Internet of Things, and specifically to a method for orderly access scheduling of charging loads, a device for orderly access scheduling of charging loads, a system for orderly access scheduling of charging loads, an edge computing terminal, a machine-readable storage medium and a computer program product. Background Art

[0002] As the number of electric vehicles connected to the grid continues to increase, the demand for charging capacity required by the distribution network increases, while the carrying capacity of the existing distribution network is limited and it is difficult to support the new high charging capacity. The existing distribution network is already facing the risk of overload of regional lines and distribution transformers and reduced stability of distribution network operation. Distributed power sources, energy storage and other resources can improve the distribution network's ability to accept electric vehicles to a certain extent. However, various charging loads, as important adjustable, controllable and flexible resources on the power consumption side, have become one of the largest loads in the power grid, with randomness, simultaneity and impact. The difficulty of maintaining safe and stable operation of the power grid through traditional means will be greatly increased. It is urgent to fully consider the load on the power consumption side and reduce the impact of charging load on the volatility of the distribution network load. Summary of the invention

[0003] The purpose of the embodiments of the present invention is to provide a method and device for orderly access scheduling of charging loads, so as to solve the problem of how to fully consider the load on the power consumption side and reduce the impact of the charging load on the fluctuation of the distribution network load.

[0004] In order to achieve the above object, an embodiment of the present invention provides a method for orderly access scheduling of charging loads, including:

[0005] Obtain real-time time series data of power load in distribution station area;

[0006] Input the real-time time series data into a load forecasting model to obtain load forecasting time series data of the distribution substation area output by the load forecasting model; wherein the load forecasting model is trained based on the historical time series data of the power load of the distribution substation area;

[0007] Receive attribute parameters of a charging load node and charging service request parameters;

[0008] Based on the load forecast time series data and the attribute parameters, calculating the access probability of the charging load node; wherein there is a negative correlation between the load forecast time series data and the access probability of the charging load node;

[0009] Determine the time of the minimum load forecast time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameter;

[0010] Based on the reserved charging load and the load forecast time series data, updated load forecast time series data of the distribution station area is calculated.

[0011] Optionally, the load prediction model is constructed based on a gated cyclic unit, and model parameters are optimized based on a Bayesian optimization method during training of the load prediction model.

[0012] Optionally, the load forecasting model is trained by the following steps:

[0013] Obtain historical time series data of power load in distribution station area;

[0014] Repeat the following steps until the set stop condition is reached:

[0015] Inputting the historical time series data into the load forecasting model to obtain forecast time series data output by the load forecasting model;

[0016] Calculating the prediction accuracy of the load prediction model based on the historical time series data and the predicted time series data, and using the prediction accuracy as an objective function;

[0017] Fitting the objective function using a Gaussian process to obtain a posterior distribution of the objective function under unevaluated hyperparameter combinations;

[0018] The hyperparameter combination of the next load forecasting model to be evaluated is selected by using the set acquisition function according to the posterior distribution of the objective function in the unevaluated hyperparameter combination.

[0019] Optionally, the attribute parameter includes the remaining energy of the charging load node; and the calculating the access probability of the charging load node based on the load forecast time series data and the attribute parameter includes:

[0020] The access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient.

[0021] Optionally, the access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient, and is calculated by the following formula:

[0022]

[0023] Among them, p k (t) represents the access probability of the charging load node, α represents the set weight coefficient, L(t) represents the load forecast time series data, LMAX Indicates the maximum value of the set load, E k represents the remaining energy of the charging load node, E full Represents the rated energy of the charging load node.

[0024] Optionally, the charging service request parameter includes charging power and charging duration, and determining the reserved charging load of the charging load node based on the charging start time and the charging service request parameter includes:

[0025] Determine a charging end time based on the charging start time and the charging duration;

[0026] The scheduled charging load of the charging load node is calculated based on the charging start time, the charging end time, and the charging power.

[0027] Optionally, the step of calculating updated load forecast time series data of a distribution station area based on the reserved charging load and the load forecast time series data includes:

[0028] Based on the sum of the reserved charging load and the load forecast time series data, updated load forecast time series data of the distribution station area is calculated.

[0029] On the other hand, an embodiment of the present invention further provides a device for orderly access scheduling of charging loads, including:

[0030] An acquisition module is used to obtain real-time time series data of power load in the distribution station area;

[0031] A prediction module, used for inputting the real-time time series data into a load prediction model to obtain load prediction time series data of the distribution substation area output by the load prediction model; wherein the load prediction model is trained based on the historical time series data of the power load of the distribution substation area;

[0032] A receiving module, used to receive attribute parameters of a charging load node and charging service request parameters;

[0033] A first calculation module, configured to calculate the access probability of the charging load node based on the load forecast time series data and the attribute parameter; wherein the load forecast time series data and the access probability of the charging load node are negatively correlated;

[0034] A second calculation module is used to determine the time of the minimum load forecast time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameter;

[0035] The third calculation module is used to calculate the updated load forecast time series data of the distribution station area based on the reserved charging load and the load forecast time series data.

[0036] Optionally, the load prediction model is constructed based on a gated cyclic unit, and model parameters are optimized based on a Bayesian optimization method during training of the load prediction model.

[0037] Optionally, the load forecasting model is trained by the following steps:

[0038] Obtain historical time series data of power load in distribution station area;

[0039] Repeat the following steps until the set stop condition is reached:

[0040] Inputting the historical time series data into the load forecasting model to obtain forecast time series data output by the load forecasting model;

[0041] Calculating the prediction accuracy of the load prediction model based on the historical time series data and the predicted time series data, and using the prediction accuracy as an objective function;

[0042] Fitting the objective function using a Gaussian process to obtain a posterior distribution of the objective function under unevaluated hyperparameter combinations;

[0043] The hyperparameter combination of the next load forecasting model to be evaluated is selected by using the set acquisition function according to the posterior distribution of the objective function in the unevaluated hyperparameter combination.

[0044] Optionally, the attribute parameter includes the remaining energy of the charging load node; and the calculating the access probability of the charging load node based on the load forecast time series data and the attribute parameter includes:

[0045] The access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient.

[0046] Optionally, the access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient, and is calculated by the following formula:

[0047]

[0048] Among them, p k (t) represents the access probability of the charging load node, α represents the set weight coefficient, L(t) represents the load forecast time series data, L MAXIndicates the maximum value of the set load, E k represents the remaining energy of the charging load node, E full Represents the rated energy of the charging load node.

[0049] Optionally, the charging service request parameter includes charging power and charging duration, and determining the reserved charging load of the charging load node based on the charging start time and the charging service request parameter includes:

[0050] Determine a charging end time based on the charging start time and the charging duration;

[0051] The scheduled charging load of the charging load node is calculated based on the charging start time, the charging end time, and the charging power.

[0052] Optionally, the step of calculating updated load forecast time series data of a distribution station area based on the reserved charging load and the load forecast time series data includes:

[0053] Based on the sum of the reserved charging load and the load forecast time series data, updated load forecast time series data of the distribution station area is calculated.

[0054] On the other hand, an embodiment of the present invention also provides an edge computing terminal, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned orderly access scheduling method of charging loads when executing the program.

[0055] On the other hand, an embodiment of the present invention further provides an orderly access scheduling system for charging loads, comprising the above-mentioned edge computing terminal, a plurality of smart meters communicatively connected to the edge computing terminal, and a plurality of charging load nodes communicatively connected to the edge computing terminal;

[0056] Each of the smart meters is used to collect real-time time series data and historical time series data of the power load in the distribution station area;

[0057] Each of the charging load nodes is used to send attribute parameters and charging service request parameters of the charging load node to the edge computing terminal;

[0058] The edge computing terminal is used to obtain real-time time series data of the power load of the distribution station area; input the real-time time series data into the load prediction model to obtain the load prediction time series data of the distribution station area output by the load prediction model; wherein the load prediction model is trained based on the historical time series data of the power load of the distribution station area; receive the attribute parameters and charging service request parameters of the charging load node; based on the load prediction time series data and the attribute parameters, calculate the access probability of the charging load node; wherein there is a negative correlation between the load prediction time series data and the access probability of the charging load node; determine the time of the minimum load prediction time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameters; based on the scheduled charging load and the load prediction time series data, calculate the updated load prediction time series data of the distribution station area.

[0059] On the other hand, the present invention also provides a machine-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned orderly access scheduling method for charging loads.

[0060] On the other hand, the present invention also provides a computer program product, including a computer program, which implements the above-mentioned orderly access scheduling method of charging loads when executed by a processor.

[0061] Through the above technical scheme, the embodiment of the present invention calculates the access probability of the charging load node based on the load forecast time series data predicted by the load forecast model and the attribute parameters of the charging load node; then determines the time of the minimum load forecast time series data corresponding to the largest access probability of the charging load node as the charging start time, and determines the reserved charging load of the charging load node based on the charging start time and the charging service request parameters; and calculates the updated load forecast time series data of the distribution station area based on the reserved charging load and the load forecast time series data to realize the orderly access scheduling of the charging load. The embodiment of the present invention realizes the orderly access scheduling of the charging load by reserving the charging resources during the low load period of the distribution station area and making full use of the period with the lowest load for charging. The present invention fully considers the load on the power consumption side and reduces the impact of the volatility of the charging load on the load of the distribution network.

[0062] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0064] Figure 1 It is a flow chart of the orderly access scheduling method of charging load provided by the present invention;

[0065] Figure 2 It is a schematic diagram of the structure of the gated recurrent unit provided by the present invention;

[0066] Figure 3 It is a schematic diagram of the training process of the gated recurrent unit based on Bayesian optimization provided by the present invention;

[0067] Figure 4 It is a structural schematic diagram of the orderly access scheduling device for charging loads provided by the present invention;

[0068] Figure 5 It is a structural schematic diagram of the edge computing terminal provided by the present invention.

[0069] Figure 6 It is a structural schematic diagram of the orderly access scheduling system of charging loads provided by the present invention. DETAILED DESCRIPTION

[0070] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0071] Method Embodiment

[0072] Please refer to Figure 1 , an embodiment of the present invention provides a method for orderly access scheduling of charging loads, comprising:

[0073] Step 100: Obtain real-time time series data of power load in the distribution station area.

[0074] The orderly access scheduling method of the charging load of the embodiment of the present invention is applied to the edge computing terminal. The edge computing terminal can obtain the real-time time series data of the power load of the distribution station area from the smart meter. Specifically, the real-time time series data of the power load of the distribution station area is the power load data of the time series. For example, it can be the power load data collected in real time by the smart meter every 15 minutes in a day. It should be noted that the real-time time series data can be the time series data of the power load of the distribution station area on the day before the forecast.

[0075] Step 200: Input the real-time time series data into a load forecasting model to obtain load forecasting time series data of the distribution station area output by the load forecasting model.

[0076] The edge computing terminal inputs the real-time time series data into the load forecasting model to obtain the load forecasting time series data of the distribution substation output by the load forecasting model. Similarly, the load forecasting time series data is the time series power load forecasting data, for example, the power load forecast of the distribution substation every 15 minutes on the forecast day. The load forecasting model is trained based on the historical time series data of the power load of the distribution substation. The historical time series data of the power load of the distribution substation can be the historical data of the power load of the distribution substation in the past period of time (for example, several months or a year).

[0077] In one embodiment, the load forecasting model can be based on various artificial intelligence models to forecast power load. For example, convolutional neural networks, deep neural networks, long short-term memory networks, and gated recurrent units can be used. In the embodiment of the present invention, please refer to Figure 2 The load forecasting model is constructed using a gated recurrent unit (GRU). The state equation of the simplified gated recurrent unit is shown in Formula 1:

[0078]

[0079] Among them, y t-1 is the hidden layer output of the previous GRU, x t is the power load data input to this round of GRU, σ is the Sigmoid function, z t is the update gate of GRU, y t ' is the updated intermediate state, y t is the output of this round of GRU; U z is the weight of the update gate, W z is the weight of the intermediate state, b z is the offset.

[0080] When the gated recurrent unit performs model training, it uses a defined loss function (such as a mean square error loss function) to calculate the loss value between the predicted result and the true value. Based on the loss value, the back propagation algorithm is used to calculate the gradient of each parameter in the model. Repeat the above steps until the model converges to complete the training of the gated recurrent unit. At this point, the load forecasting model based on the gated recurrent unit can input real-time time series data to obtain the load forecasting time series data of the distribution station area.

[0081] In other embodiments, in order to reduce the operational complexity of edge computing terminals and improve the computational efficiency of load forecasting by edge computing terminals. The embodiment of the present invention constructs a load forecasting model of a gated cyclic unit based on Bayesian Optimization (BO). The power load is predicted by the gated cyclic unit, and the model parameters of the gated cyclic unit are fine-tuned in combination with the Bayesian optimization method. It should be noted that in the load forecasting model of the gated cyclic unit based on Bayesian Optimization (BO), the structure of the gated cyclic unit is similar to Figure 2 The same, no further elaboration here.

[0082] Please refer to Figure 3 The load forecasting model is trained by optimizing the model parameters based on the Bayesian optimization method through the following steps:

[0083] Step 10: Obtain the historical time series data of the power load in the distribution station area.

[0084] Repeat the following steps until the set stop condition is reached:

[0085] Step 20: Input the historical time series data into the load forecasting model to obtain the forecast time series data output by the load forecasting model.

[0086] Step 30: Calculate the prediction accuracy of the load prediction model based on the historical time series data and the predicted time series data, and use the prediction accuracy as the objective function.

[0087] Step 40: Fit the objective function using a Gaussian process to obtain the posterior distribution of the objective function under unevaluated hyperparameter combinations.

[0088] Step 50: Select the next hyperparameter combination of the load forecasting model to be evaluated according to the posterior distribution of the objective function in the unevaluated hyperparameter combination by using the set acquisition function.

[0089] First, obtain the historical data of the power load of the distribution station area in the past period of time (for example, several months or a year). The edge computing terminal can pre-process the historical time series data of the power load of the distribution station area by data cleaning and data normalization. Among them, data cleaning includes processing missing values, outliers and duplicate data to ensure the accuracy and integrity of the data. Normalize the data to improve the training efficiency and prediction accuracy of the model. In an embodiment of the present invention, the historical time series data of the power load of the distribution station area can also be divided into a training data set and a test data set for model training and testing. Then determine the initial model parameters, such as the number of GRU units, the number of hidden units, the learning rate and the regularization parameter. In an embodiment of the present invention, the number of GRU units, the number of hidden units, the learning rate and the regularization parameter are collectively referred to as a hyperparameter combination. The edge computing terminal inputs the historical time series data into the load forecasting model to obtain the predicted time series data output by the load forecasting model. The prediction accuracy (for example, mean square error) of the load forecasting model is calculated based on the historical time series data and the predicted time series data, and the prediction accuracy is used as the objective function.

[0090] The Bayes-GRU prediction model of the embodiment of the present invention introduces Bayesian optimization to optimize and adjust the model parameters of the gated recurrent unit. The optimization target is shown in Formula-2, where s is a hyperparameter combination of the gated recurrent unit (such as the number of units in the hidden layer, the learning rate), S is a hyperparameter set, f(s) is a mapping of the hyperparameter combination s to the model generalization performance, and s opt is the optimal hyperparameter combination.

[0091]

[0092] From Bayes' theorem we know that:

[0093]

[0094] Among them, GP is the Gaussian distribution function, ∑(s 1:t ,s 1:t ) is the covariance matrix, f(s t+1 ) is the mapping function from the t+1th optimized hyperparameter combination s to the model generalization performance, f(s 1:t ) is the mapping function of the hyperparameter combination obtained from the first to t optimizations, u(s 1:t ) is the mean value of the hyperparameter combination optimized from 1 to t times. Through continuous iterative updates, s opt =s t+1 , and finally get the most optimized hyperparameter combination of gated recurrent units. P(f(s t+1 )|f(s 1:t )) indicates that in f(s 1:t ) occurs when f(s t+1) occurs, P(f(s 1:t )|f(s t+1 )) indicates that in f(s t+1 ) occurs when f(s 1:t ) occurs, P(f(s t+1 )) represents f(s t+1 ) is the prior probability or marginal probability of .

[0095] The edge computing terminal builds a surrogate model, that is, a probability model such as Gaussian Process (GP) is used to approximate the objective function. The Gaussian process can predict the performance of unknown hyperparameter combinations based on known hyperparameter combinations and their corresponding model performance, and give the corresponding confidence interval, that is, obtain the posterior distribution of the objective function in the unevaluated hyperparameter combination. Then, an acquisition function is defined to select the next hyperparameter combination to be evaluated based on the posterior distribution of the surrogate model. Common acquisition functions include: Expected Improvement (EI), Probability of Improvement (PI), Upper Confidence Bound (UCB), etc. These functions usually strike a balance between exploration and exploitation. Repeat the above steps, continuously update the surrogate model and select new hyperparameter combinations for evaluation until the preset stop condition is reached, such as reaching the maximum number of iterations or the objective function value meets the requirements.

[0096] Thus, the edge computing terminal collects the historical time series data of all smart meters in the entire distribution station area and stores it in the database. The historical time series data is divided into a training set and a test set, the gated cyclic unit is trained, and the model parameters of the gated cyclic unit are optimized using the Bayesian optimization method to form a complete Bayes-GRU load prediction model. Finally, the Bayes-GRU load prediction model is called to form the load prediction time series data L(t) of the distribution station area. The embodiment of the present invention combines a lightweight gated cyclic unit with Bayesian optimization through an electricity load prediction method constructed based on a Bayesian optimization-gated cyclic unit (Bayes-GRU) model, thereby reducing the operating complexity of the edge computing terminal and improving the computational efficiency of the edge computing terminal for load prediction.

[0097] Step 300: Receive attribute parameters of a charging load node and charging service request parameters.

[0098] The edge computing terminal receives a charging service request from a charging load node k. The charging service request includes attribute parameters of the charging load node and charging service request parameters. The attribute parameters of the charging load node may represent the parameters of the charging load node's own performance, such as the remaining energy E of the charging load node. k , Rated energy E of charging load node full The charging service request parameter indicates the parameters related to the charging request, such as the charging power L C 、Charging time T C wait.

[0099] Step 400: Calculate the access probability of the charging load node based on the load forecast time series data and the attribute parameters.

[0100] The edge computing terminal calculates the access probability of the charging load node based on the load forecast time series data and the attribute parameters. The load forecast time series data is negatively correlated with the access probability of the charging load node. That is, the smaller the value of the load forecast time series data, the greater the access probability of the charging load node. Therefore, when the access probability of the charging load node is the largest, the value of the corresponding load forecast time series data is the smallest, and this is the period when the load of the distribution station area is the lowest.

[0101] In one embodiment, the access probability of the charging load node is calculated based on the load forecast time series data and the attribute parameters, including: calculating the access probability of the charging load node based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set maximum load value and the set weight coefficient.

[0102] Specifically, the access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient, and is calculated by the following formula:

[0103]

[0104] Among them, p k (t) represents the access probability of the charging load node, α represents the set weight coefficient, L(t) represents the load forecast time series data, L MAX Indicates the maximum value of the set load, E k represents the remaining energy of the charging load node, E full Represents the rated energy of the charging load node.

[0105] It can be seen that in Formula 4, there is a negative correlation between the load forecast time series data and the access probability of the charging load node. Therefore, when the access probability of the charging load node is the largest, the value of the corresponding load forecast time series data is the smallest, and this is the time period when the load of the distribution station area is at its lowest. By determining the time of the minimum load forecast time series data corresponding to the largest access probability of the charging load node as the charging start time, the embodiment of the present invention schedules the charging load to prioritize the time period when the load of the distribution station area is at its lowest, so as to improve the stability of the charging load operation in the distribution station area.

[0106] Step 500: determine the time of the minimum load forecast time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameters.

[0107] The edge computing terminal determines the time of the minimum load forecast time series data corresponding to the access probability of the maximum charging load node as the charging start time, which is expressed by the following formula: Among them, p k (t) represents the access probability of the charging load node, t 1 Indicates the charging start time, T day It represents a time set from 0:00 to 24:00 in a day. The edge computing terminal then determines the scheduled charging load of the charging load node based on the charging start time, charging power and charging duration.

[0108] In one embodiment, determining the scheduled charging load of the charging load node based on the charging start time and the charging service request parameters includes: determining the charging end time based on the charging start time and the charging duration; and calculating the scheduled charging load of the charging load node based on the charging start time, the charging end time and the charging power.

[0109] The charging end time is determined based on the charging start time and the charging duration, and is calculated by the following formula: 2 =t 1 +T C ; where t 1 Indicates the charging start time, T C Indicates the charging time, t 2 Indicates the charging end time. Based on the charging start time, the charging end time and the charging power, the scheduled charging load of the charging load node is calculated by the following formula:

[0110] where t 1 Indicates the charging start time, t 2 Indicates the charging end time, LC Indicates the charging power.

[0111] The embodiment of the present invention schedules the charging load to charge in the time period with the lowest load in the distribution station area according to the peak and valley conditions of the power load, so as to improve the stability of the charging load operation in the distribution station area.

[0112] Step 600: Calculate updated load forecast time series data of the distribution station area based on the reserved charging load and the load forecast time series data.

[0113] The edge computing terminal calculates updated load forecast time series data based on the scheduled charging load and the load forecast time series data, and completes the reservation and control of the charging load access. In one embodiment, the calculation of updated load forecast time series data of the distribution station area based on the scheduled charging load and the load forecast time series data includes: calculating updated load forecast time series data of the distribution station area based on the sum of the scheduled charging load and the load forecast time series data.

[0114] Specifically, the updated load forecast time series data is calculated according to Formula-5:

[0115]

[0116] Among them, t 1 Indicates the charging start time, T C Indicates the charging time, t 2 represents the charging end time, L(t) represents the load forecast time series data, and L(t)' represents the updated load forecast time series data.

[0117] The embodiment of the present invention calculates the access probability of the charging load node based on the load forecast time series data predicted by the load forecast model and the attribute parameters of the charging load node; then determines the time of the minimum load forecast time series data corresponding to the largest access probability of the charging load node as the charging start time, and determines the reserved charging load of the charging load node based on the charging start time and the charging service request parameters; and calculates the updated load forecast time series data of the distribution station area based on the reserved charging load and the load forecast time series data to achieve orderly access scheduling of the charging load. The embodiment of the present invention makes full use of the time period with the lowest load to charge by reserving charging resources during the low load period of the distribution station area to achieve orderly access scheduling of the charging load. The present invention fully considers the load on the power consumption side and reduces the impact of the volatility of the charging load on the load of the distribution network.

[0118] In summary, the embodiment of the present invention constructs a prediction method for load prediction time series data of a distribution station area based on the Bayes-GRU model, combines a lightweight gated recurrent unit with a Bayesian optimization parameter, and reduces the operating complexity of its edge computing terminal; and through the reservation access control mechanism of the charging load access probability of the distribution station area, full use is made of the period of lowest load for charging, thereby realizing orderly access scheduling of the charging load.

[0119] Device Embodiment

[0120] Please refer to Figure 4 On the other hand, an embodiment of the present invention further provides a device for orderly access scheduling of charging loads, including:

[0121] The acquisition module 401 is used to acquire the real-time time series data of the power load of the distribution station area;

[0122] The prediction module 402 is used to input the real-time time series data into the load prediction model to obtain the load prediction time series data of the distribution substation output by the load prediction model; wherein the load prediction model is trained based on the historical time series data of the power load of the distribution substation;

[0123] The receiving module 403 is used to receive the attribute parameters of the charging load node and the charging service request parameters;

[0124] A first calculation module 404 is used to calculate the access probability of the charging load node based on the load forecast time series data and the attribute parameter; wherein there is a negative correlation between the load forecast time series data and the access probability of the charging load node;

[0125] The second calculation module 405 is used to determine the time of the minimum load forecast time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameter;

[0126] The third calculation module 406 is used to calculate the updated load forecast time series data of the distribution station area based on the reserved charging load and the load forecast time series data.

[0127] The embodiment of the present invention calculates the access probability of the charging load node based on the load forecast time series data predicted by the load forecast model and the attribute parameters of the charging load node; then determines the time of the minimum load forecast time series data corresponding to the largest access probability of the charging load node as the charging start time, and determines the reserved charging load of the charging load node based on the charging start time and the charging service request parameters; and calculates the updated load forecast time series data of the distribution station area based on the reserved charging load and the load forecast time series data to achieve orderly access scheduling of the charging load. The embodiment of the present invention makes full use of the time period with the lowest load to charge by reserving charging resources during the low load period of the distribution station area to achieve orderly access scheduling of the charging load. The present invention fully considers the load on the power consumption side and reduces the impact of the volatility of the charging load on the load of the distribution network.

[0128] Optionally, the load prediction model is constructed based on a gated cyclic unit, and model parameters are optimized based on a Bayesian optimization method during training of the load prediction model.

[0129] Optionally, the load forecasting model is trained by the following steps:

[0130] Obtain historical time series data of power load in distribution station area;

[0131] Repeat the following steps until the set stop condition is reached:

[0132] Inputting the historical time series data into the load forecasting model to obtain forecast time series data output by the load forecasting model;

[0133] Calculating the prediction accuracy of the load prediction model based on the historical time series data and the predicted time series data, and using the prediction accuracy as an objective function;

[0134] Fitting the objective function using a Gaussian process to obtain a posterior distribution of the objective function under unevaluated hyperparameter combinations;

[0135] The hyperparameter combination of the next load forecasting model to be evaluated is selected by using the set acquisition function according to the posterior distribution of the objective function in the unevaluated hyperparameter combination.

[0136] Optionally, the attribute parameter includes the remaining energy of the charging load node; and the calculating the access probability of the charging load node based on the load forecast time series data and the attribute parameter includes:

[0137] The access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient.

[0138] Optionally, the access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient, and is calculated by the following formula:

[0139]

[0140] Among them, p k (t) represents the access probability of the charging load node, α represents the set weight coefficient, L(t) represents the load forecast time series data, L MAX Indicates the maximum value of the set load, E k represents the remaining energy of the charging load node, E full Represents the rated energy of the charging load node.

[0141] Optionally, the charging service request parameter includes charging power and charging duration, and determining the reserved charging load of the charging load node based on the charging start time and the charging service request parameter includes:

[0142] Determine a charging end time based on the charging start time and the charging duration;

[0143] The scheduled charging load of the charging load node is calculated based on the charging start time, the charging end time, and the charging power.

[0144] Optionally, the step of calculating updated load forecast time series data of a distribution station area based on the reserved charging load and the load forecast time series data includes:

[0145] Based on the sum of the reserved charging load and the load forecast time series data, updated load forecast time series data of the distribution station area is calculated.

[0146] The orderly access scheduling device for charging loads includes a processor and a memory. The acquisition module 401, prediction module 402, receiving module 403, first calculation module 404, second calculation module 405 and third calculation module 406 are all stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.

[0147] The processor includes a kernel, which calls the corresponding program unit from the memory. There can be one or more kernels.

[0148] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0149] Figure 5 The following is a schematic diagram of the physical structure of an edge computing terminal. Figure 5 As shown, the edge computing terminal may include: a processor (processor) 510, a communication interface (Communications Interface) 520, a memory (memory) 530 and a communication bus 540, wherein the processor 510, the communication interface 520, and the memory 530 communicate with each other through the communication bus 540. The processor 510 can call the logic instructions in the memory 530 to execute the orderly access scheduling method of the charging load, the method including: obtaining real-time time series data of the power load of the distribution station area; inputting the real-time time series data into the load prediction model to obtain the load prediction time series data of the distribution station area output by the load prediction model; wherein the load prediction model is trained based on the historical time series data of the power load of the distribution station area; receiving the attribute parameters and charging service request parameters of the charging load node; based on the load prediction time series data and the attribute parameters, calculating the access probability of the charging load node; wherein there is a negative correlation between the load prediction time series data and the access probability of the charging load node; determining the time of the minimum load prediction time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determining the reserved charging load of the charging load node based on the charging start time and the charging service request parameters; based on the reserved charging load and the load prediction time series data, calculating the updated load prediction time series data of the distribution station area.

[0150] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0151] On the other hand, please refer to Figure 6The embodiment of the present invention further provides an orderly access scheduling system for charging loads, including the above-mentioned edge computing terminal 10, a plurality of smart meters 20 communicatively connected to the edge computing terminal 10, and a plurality of charging load nodes communicatively connected to the edge computing terminal 10. The smart meter 20 (i.e., meter 1 to meter N in the figure) can be communicatively connected to the smart meter 20 via HPLC or a wireless communication network. The charging load node can be a charging pile 30. The charging pile 30 can be communicatively connected to the smart meter 20 via HPLC or a wireless communication network.

[0152] Each of the smart meters 20 is used to collect real-time time series data and historical time series data of the power load in the distribution station area; each of the charging load nodes is used to send the attribute parameters and charging service request parameters of the charging load node to the edge computing terminal 10.

[0153] The edge computing terminal 10 is used to obtain real-time time series data of the power load in the distribution station area; input the real-time time series data into the load prediction model to obtain the load prediction time series data of the distribution station area output by the load prediction model; wherein the load prediction model is trained based on the historical time series data of the power load in the distribution station area; receive the attribute parameters and charging service request parameters of the charging load node; based on the load prediction time series data and the attribute parameters, calculate the access probability of the charging load node; wherein there is a negative correlation between the load prediction time series data and the access probability of the charging load node; determine the time of the minimum load prediction time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameters; based on the scheduled charging load and the load prediction time series data, calculate the updated load prediction time series data of the distribution station area.

[0154] The edge computing terminal 10 mainly runs the following functional APPs:

[0155] (1) Electricity usage information collection APP: collects historical and real-time time series data of all smart meters 20 in the distribution area and stores them in the database.

[0156] (2) Power load forecasting APP: Train the Bayes-GRU model to form a complete forecasting model. Then run the load forecasting model based on Bayes-GRU to forecast the power load of the distribution station area on the same day based on the historical time series data and real-time time series data of the power load.

[0157] (3) Orderly charging scheduling APP: Combined with the load forecast time series data, the access probability of each charging load is calculated to make reservations and access control of charging resources.

[0158] The embodiment of the present invention is based on a Bayesian optimized GRU power load prediction model to predict the dynamic rechargeable capacity of a distribution station area. It also uses a reservation-based access control method for charging load access probability to schedule the charging load to prioritize charging during the period when the load in the distribution station area is at its lowest, according to the peak and valley conditions of the power load, so as to improve the stability of the charging load operation in the distribution station area.

[0159] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a machine-readable storage medium. When the computer program is executed by a processor, the computer can execute an orderly access scheduling method for charging loads, the method including: obtaining real-time time series data of the power load in the distribution station area; inputting the real-time time series data into a load prediction model to obtain the load prediction time series data of the distribution station area output by the load prediction model; wherein the load prediction model is trained based on the historical time series data of the power load in the distribution station area; receiving attribute parameters and charging service request parameters of a charging load node; calculating the access probability of the charging load node based on the load prediction time series data and the attribute parameters; wherein there is a negative correlation between the load prediction time series data and the access probability of the charging load node; determining the time of the minimum load prediction time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determining the reserved charging load of the charging load node based on the charging start time and the charging service request parameters; calculating the updated load prediction time series data of the distribution station area based on the reserved charging load and the load prediction time series data.

[0160] On the other hand, the present invention also provides a machine-readable storage medium having a computer program stored thereon, which is implemented when executed by a processor to execute an orderly access scheduling method for charging loads, the method comprising: obtaining real-time time series data of the power load in a distribution station area; inputting the real-time time series data into a load prediction model to obtain load prediction time series data of the distribution station area output by the load prediction model; wherein the load prediction model is trained based on historical time series data of the power load in the distribution station area; receiving attribute parameters and charging service request parameters of a charging load node; calculating the access probability of the charging load node based on the load prediction time series data and the attribute parameters; wherein there is a negative correlation between the load prediction time series data and the access probability of the charging load node; determining the time of the minimum load prediction time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determining the reserved charging load of the charging load node based on the charging start time and the charging service request parameters; calculating the updated load prediction time series data of the distribution station area based on the reserved charging load and the load prediction time series data.

[0161] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0162] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for orderly access scheduling of charging loads, characterized in that: include: Obtain real-time time series data of power load in distribution station area; Input the real-time time series data into a load forecasting model to obtain load forecasting time series data of the distribution substation area output by the load forecasting model; wherein the load forecasting model is trained based on the historical time series data of the power load of the distribution substation area; Receive attribute parameters of a charging load node and charging service request parameters; Based on the load forecast time series data and the attribute parameters, calculating the access probability of the charging load node; wherein there is a negative correlation between the load forecast time series data and the access probability of the charging load node; Determine the time of the minimum load forecast time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameter; Based on the reserved charging load and the load forecast time series data, updated load forecast time series data of the distribution station area is calculated.

2. The method for orderly access and scheduling of charging loads according to claim 1, characterized in that: The load prediction model is constructed based on a gated cyclic unit, and model parameter optimization is performed based on a Bayesian optimization method during training of the load prediction model.

3. The method for orderly access and scheduling of charging loads according to claim 2, characterized in that: The load forecasting model is trained through the following steps: Obtain historical time series data of power load in distribution station area; Repeat the following steps until the set stop condition is reached: Inputting the historical time series data into the load forecasting model to obtain forecast time series data output by the load forecasting model; Calculating the prediction accuracy of the load prediction model based on the historical time series data and the predicted time series data, and using the prediction accuracy as an objective function; Fitting the objective function using a Gaussian process to obtain a posterior distribution of the objective function under unevaluated hyperparameter combinations; The hyperparameter combination of the next load forecasting model to be evaluated is selected by using the set acquisition function according to the posterior distribution of the objective function in the unevaluated hyperparameter combination.

4. The method for orderly access and scheduling of charging loads according to claim 1, characterized in that: The attribute parameter includes the remaining energy of the charging load node; and the calculating the access probability of the charging load node based on the load forecast time series data and the attribute parameter includes: The access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient.

5. The method for orderly access and scheduling of charging loads according to claim 4, characterized in that: The access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient, and is calculated by the following formula: Among them, p k (t) represents the access probability of the charging load node, α represents the set weight coefficient, L(t) represents the load forecast time series data, L MAX Indicates the maximum value of the set load, E k represents the remaining energy of the charging load node, E full Represents the rated energy of the charging load node.

6. The method for orderly access and scheduling of charging loads according to claim 1, characterized in that: The charging service request parameters include charging power and charging duration, and determining the scheduled charging load of the charging load node based on the charging start time and the charging service request parameters includes: Determine a charging end time based on the charging start time and the charging duration; The scheduled charging load of the charging load node is calculated based on the charging start time, the charging end time, and the charging power.

7. The method for orderly access and scheduling of charging loads according to claim 1, characterized in that: The step of calculating updated load forecast time series data of the distribution station area based on the scheduled charging load and the load forecast time series data includes: Based on the sum of the reserved charging load and the load forecast time series data, updated load forecast time series data of the distribution station area is calculated.

8. An orderly access scheduling device for charging loads, characterized in that: include: An acquisition module is used to obtain real-time time series data of power load in the distribution station area; A prediction module, used for inputting the real-time time series data into a load prediction model to obtain load prediction time series data of the distribution substation area output by the load prediction model; wherein the load prediction model is trained based on the historical time series data of the power load of the distribution substation area; A receiving module, used to receive attribute parameters of a charging load node and charging service request parameters; A first calculation module, configured to calculate the access probability of the charging load node based on the load forecast time series data and the attribute parameter; wherein the load forecast time series data and the access probability of the charging load node are negatively correlated; A second calculation module is used to determine the time of the minimum load forecast time series data corresponding to the maximum access probability of the charging load node as the charging start time, and determine the scheduled charging load of the charging load node based on the charging start time and the charging service request parameter; The third calculation module is used to calculate the updated load forecast time series data of the distribution station area based on the reserved charging load and the load forecast time series data.

9. The orderly access scheduling device for charging load according to claim 8, characterized in that: The load prediction model is constructed based on a gated cyclic unit, and model parameter optimization is performed based on a Bayesian optimization method during training of the load prediction model.

10. The orderly access scheduling device for charging load according to claim 9, characterized in that: The load forecasting model is trained through the following steps: Obtain historical time series data of power load in distribution station area; Repeat the following steps until the set stop condition is reached: Inputting the historical time series data into the load forecasting model to obtain forecast time series data output by the load forecasting model; Calculating the prediction accuracy of the load prediction model based on the historical time series data and the predicted time series data, and using the prediction accuracy as an objective function; Fitting the objective function using a Gaussian process to obtain a posterior distribution of the objective function under unevaluated hyperparameter combinations; The hyperparameter combination of the next load forecasting model to be evaluated is selected by using the set acquisition function according to the posterior distribution of the objective function in the unevaluated hyperparameter combination.

11. The orderly access scheduling device for charging load according to claim 8, characterized in that: The attribute parameter includes the remaining energy of the charging load node; and the calculating the access probability of the charging load node based on the load forecast time series data and the attribute parameter includes: The access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient.

12. The orderly access scheduling device for charging load according to claim 11, characterized in that: The access probability of the charging load node is calculated based on the load forecast time series data, the remaining energy of the charging load node, the rated energy of the charging load node, the set load maximum value and the set weight coefficient, and is calculated by the following formula: Among them, p k (t) represents the access probability of the charging load node, α represents the set weight coefficient, L(t) represents the load forecast time series data, L MAX Indicates the maximum value of the set load, E k represents the remaining energy of the charging load node, E full Represents the rated energy of the charging load node.

13. The orderly access scheduling device for charging load according to claim 8, characterized in that: The charging service request parameters include charging power and charging duration, and determining the scheduled charging load of the charging load node based on the charging start time and the charging service request parameters includes: Determine a charging end time based on the charging start time and the charging duration; The scheduled charging load of the charging load node is calculated based on the charging start time, the charging end time, and the charging power.

14. The orderly access scheduling device for charging load according to claim 8, characterized in that: The step of calculating updated load forecast time series data of the distribution station area based on the scheduled charging load and the load forecast time series data includes: Based on the sum of the reserved charging load and the load forecast time series data, updated load forecast time series data of the distribution station area is calculated.

15. An edge computing terminal, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the orderly access scheduling method of charging loads described in any one of claims 1 to 7 is implemented.

16. An orderly access dispatching system for charging loads, characterized in that: The device comprises the edge computing terminal as claimed in claim 15, a plurality of smart meters communicatively connected to the edge computing terminal, and a plurality of charging load nodes communicatively connected to the edge computing terminal; Each of the smart meters is used to collect real-time time series data and historical time series data of the power load in the distribution station area; Each of the charging load nodes is used to send attribute parameters and charging service request parameters of the charging load node to the edge computing terminal; The edge computing terminal is used to obtain real-time time series data of power load in the distribution station area; The real-time time series data is input into the load forecasting model to obtain the load forecasting time series data of the distribution station area output by the load forecasting model; wherein the load forecasting model is trained based on the historical time series data of the power load in the distribution station area; the attribute parameters and charging service request parameters of the charging load node are received; based on the load forecasting time series data and the attribute parameters, the access probability of the charging load node is calculated; wherein there is a negative correlation between the load forecasting time series data and the access probability of the charging load node; the time of the minimum load forecasting time series data corresponding to the maximum access probability of the charging load node is determined as the charging start time, and the scheduled charging load of the charging load node is determined based on the charging start time and the charging service request parameters; based on the scheduled charging load and the load forecasting time series data, the updated load forecasting time series data of the distribution station area is calculated.

17. A machine-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for orderly access scheduling of charging loads according to any one of claims 1 to 7 is implemented.

18. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for orderly access scheduling of charging loads according to any one of claims 1 to 7 is implemented.

Images