Probability load prediction method and system for urban and rural residents

Abstract

The invention relates to a probability load prediction method and system for urban and rural residents. The method comprises the following steps: acquiring historical load data of the urban and rural residents and preprocessing the historical load data; the preprocessed historical load data serve as a sample set, a prediction model based on a neural network is trained, the prediction model takes a probability load interval as model output, a quantile loss function is constructed through a hyperbolic cosine function, iteration training is conducted on the prediction model based on the quantile loss function, and a prediction result is obtained; obtaining a trained probability load prediction model; and carrying out probability load prediction on urban and rural residents by utilizing the trained probability load prediction model.

Description

Technical Field

[0001] The invention relates to a probabilistic load forecasting method and system for urban and rural residents, belonging to the technical field of power load forecasting. Background Art

[0002] Power load forecasting is an important part of power system planning and operation, and accurate forecasting of power load is crucial.

[0003] At present, most traditional load forecasts are deterministic point forecasts, that is, forecasts are made based on the user's historical electricity consumption data, and the predicted value of the load is finally obtained. However, because the proportion of uncertain loads in the power system has increased dramatically, and the electricity consumption of urban and rural residents is easily affected by weather and social events, traditional point forecasts can no longer support their variability and uncertainty. In recent years, load probability forecasting has become increasingly important. Most of the early research focused on probabilistic load forecasting using QRNN and LSTM. One disadvantage of QRNN is that the computational time for training the model is long, which increases the training cost of large data sets. Another problem with the improved QRNN is overfitting. In short, there is still room for further improvement in the prediction accuracy of the probabilistic load forecasting model. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention proposes a probabilistic load forecasting method and system for urban and rural residents.

[0005] The technical solution of the present invention is as follows:

[0006] On the one hand, the present invention proposes a probabilistic load forecasting method for urban and rural residents, comprising the following steps:

[0007] Obtain historical load data of urban and rural residents and perform preprocessing;

[0008] The preprocessed historical load data is used as a sample set to train a prediction model based on a neural network. The prediction model uses a probabilistic load interval as a model output, and a quantile loss function is constructed through a hyperbolic cosine function. The prediction model is iteratively trained based on the quantile loss function to obtain a trained probabilistic load prediction model.

[0009] The trained probabilistic load forecasting model is used to carry out probabilistic load forecasting for urban and rural residents.

[0010] As a preferred implementation, the step of using the pre-processed historical load data as a sample set to train a prediction model based on a neural network includes:

[0011] For the historical load data in the sample set, the time characteristics, meteorological characteristics and user characteristics in the historical load data are represented as numerical data in a one-hot encoding manner and then input into the prediction model.

[0012] As a preferred implementation, the prediction model uses an LSTM neural network including an encoder and a decoder.

[0013] As a preferred implementation, the quantile loss function is specifically:

[0014]

[0015] Among them, q is the regression loss, which ranges from 0 to 1, and y t is the true value, is the predicted value at the quantile at time t, and a is the parameter for adjusting the smoothness.

[0016] On the other hand, the present invention also proposes a probabilistic load forecasting system for urban and rural residents, comprising:

[0017] A data acquisition module, used to acquire historical load data of urban and rural residents and perform preprocessing;

[0018] A model training module is used to train a prediction model based on a neural network using preprocessed historical load data as a sample set. The prediction model uses a probabilistic load interval as a model output, and constructs a quantile loss function through a hyperbolic cosine function. The prediction model is iteratively trained based on the quantile loss function to obtain a trained probabilistic load prediction model.

[0019] The probabilistic load forecasting module uses the trained probabilistic load forecasting model to perform probabilistic load forecasting for urban and rural residents.

[0020] As a preferred implementation, the step of using the pre-processed historical load data as a sample set to train a prediction model based on a neural network includes:

[0021] For the historical load data in the sample set, the time characteristics, meteorological characteristics and user characteristics in the historical load data are represented as numerical data in a one-hot encoding manner and then input into the prediction model.

[0022] As a preferred implementation, the prediction model uses an LSTM neural network including an encoder and a decoder.

[0023] As a preferred implementation, the quantile loss function is specifically:

[0024]

[0025] Among them, y t is the true value, is the predicted value at the quantile at time t, and a is the parameter for adjusting the smoothness.

[0026] On the other hand, the present invention further proposes an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any embodiment of the present invention when executing the program.

[0027] On the other hand, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any embodiment of the present invention.

[0028] Additional aspects and advantages of the present invention will be set forth in the following description, and some of them will be apparent from the description, or may be understood by practicing the present invention. In addition, the various aspects and advantages of the present invention may be realized and obtained by the method steps and combinations particularly pointed out in the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 This is a schematic diagram of a method flow chart of Embodiment 1 of the present invention;

[0030] Figure 2 Schematic diagram of the working principle of the encoder-decoder in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] 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.

[0032] It should be understood that the step numbers used in this document are only for convenience of description and are not intended to limit the order in which the steps are executed.

[0033] It should be understood that the terms used in the present specification are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the present specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0034] The terms “include” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0035] The term "and / or" means and includes any and all possible combinations of one or more of the associated listed items.

[0036] Embodiment 1:

[0037] In order to solve the problem of how to better consider the low prediction accuracy caused by the increase in the proportion of uncertain loads in the power system. This embodiment uses the log-cosh loss function as the loss function of the prediction model, which is more robust to abnormal points in the data and has a variable gradient; the prediction model with an encoder-decoder structure is used for prediction to improve the prediction flexibility.

[0038] See also Figure 1 The present embodiment provides a probabilistic load forecasting method for urban and rural residents, which specifically includes the following steps:

[0039] S100, obtaining historical load data of urban and rural residents and performing preprocessing; the preprocessing methods include normalization and filling missing values ​​with linear interpolation.

[0040] S200, using the preprocessed historical load data as a sample set, training a prediction model based on a neural network, wherein the prediction model uses a probabilistic load interval as a model output, and constructs a quantile loss function through a hyperbolic cosine function, and iteratively trains the prediction model based on the quantile loss function to obtain a trained probabilistic load prediction model;

[0041] S300, using the trained probabilistic load forecasting model to perform probabilistic load forecasting for urban and rural residents.

[0042] As a preferred implementation of this embodiment, the step of using the pre-processed historical load data as a sample set to train a prediction model based on a neural network includes:

[0043] For the historical load data in the sample set, the time characteristics, meteorological characteristics and user characteristics in the historical load data are represented as numerical data in a one-hot encoding manner and then input into the prediction model. Among them, the user characteristics mainly include the user's age, employment status, income, family population, etc.

[0044] like Figure 2 As shown, as a preferred implementation of this embodiment, the prediction model adopts an LSTM neural network including an encoder and a decoder. The encoder part is stacked with the LSTM unit, and each unit allows a single element in the load sequence as input.

[0045] The decoder generates output at each LSTM unit. The information of all input elements is forwarded to the decoder for prediction. The mathematical equations of the encoder-decoder model are shown in (1) and (2):

[0046] h t =LSTM encoder (x t ,h t-1 ) (1)

[0047] h′ t =LSTM decoder (y t ,h′ t-1 ) (2)

[0048] In the formula, x t ={x1,x2,...,x T} and y t ={y1,y2,...,y T} are the input time series and target output sequence of the prediction model respectively. t is the encoder state vector, i.e., the final hidden state produced by the encoder, and h t It also serves as the initial state of the decoder part. To determine the output, in each LSTM unit, the state of the previous layer and the output of the last LSTM unit are used as input to calculate the output of the next hidden state. t represents the final output at time step t.

[0049] The encoder-decoder based LSTM network can outline sequences of different lengths from each other, that is, the sizes of input and output sequences can be different, which improves the flexibility of prediction.

[0050] As a preferred implementation of this embodiment, the quantile loss function uses log-cosh, which is a hyperbolic cosine function based on the error value. The log-cosh loss function is an estimator that selects the relative solution of the median rather than the average, which helps to find a better balance between reconstruction and latent space optimization, thereby improving accuracy performance. The log-cosh loss function is shown in equations (3) and (4):

[0051]

[0052] Where a is a parameter for adjusting the smoothness. Since the log-cosh loss function is robust to outliers and has a variable gradient.

[0053] In order to perform probability prediction, this embodiment changes the function to:

[0054]

[0055] Among them, q is the regression loss, which ranges from 0 to 1, and y t is the true value, is the predicted value at the quantile at time t.

[0056] Based on the above implementation, the probabilistic load forecasting method for urban and rural residents provided in this embodiment utilizes the log-cosh loss function, which is more tolerant to outliers in the sample set. Combined with the LSTM model, it can improve the accuracy of power load forecasting and better quantify uncertainty.

[0057] Embodiment 2:

[0058] This embodiment proposes a probabilistic load forecasting system for urban and rural residents, including:

[0059] A data acquisition module, used to acquire historical load data of urban and rural residents and perform preprocessing; this module is used to implement the function of step S100 in the first embodiment, which will not be described in detail here;

[0060] A model training module is used to train a prediction model based on a neural network using preprocessed historical load data as a sample set, wherein the prediction model uses a probabilistic load interval as a model output, and constructs a quantile loss function through a hyperbolic cosine function, and iteratively trains the prediction model based on the quantile loss function to obtain a trained probabilistic load prediction model; this module is used to implement the function of step S200 in the first embodiment, and will not be described in detail here;

[0061] The probabilistic load forecasting module uses the trained probabilistic load forecasting model to perform probabilistic load forecasting for urban and rural residents; this module is used to implement the function of step S300 in the first embodiment, which will not be described in detail here.

[0062] As a preferred implementation of this embodiment, the step of using the pre-processed historical load data as a sample set to train a prediction model based on a neural network includes:

[0063] For the historical load data in the sample set, the time characteristics, meteorological characteristics and user characteristics in the historical load data are represented as numerical data in a one-hot encoding manner and then input into the prediction model.

[0064] As a preferred implementation of this embodiment, the prediction model adopts an LSTM neural network including an encoder and a decoder.

[0065] As a preferred implementation of this embodiment, the quantile loss function is specifically:

[0066]

[0067] Among them, y t is the true value, is the predicted value at the quantile at time t.

[0068] Embodiment three:

[0069] This embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any embodiment of the present invention when executing the program.

[0070] Embodiment 4:

[0071] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method described in any embodiment of the present invention is implemented.

[0072] In the embodiments of the present application, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can be represented by: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c can be single or multiple.

[0073] Those of ordinary skill in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented in a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0074] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0075] In several embodiments provided in the present application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, 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, which is stored in a storage medium and includes several 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 application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), disk or optical disk, and other media that can store program codes.

[0076] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A probabilistic load forecasting method for urban and rural residents, characterized in that: The following steps are involved: Obtain historical load data of urban and rural residents and perform preprocessing; The preprocessed historical load data is used as a sample set to train a prediction model based on a neural network. The prediction model uses a probabilistic load interval as a model output, and a quantile loss function is constructed through a hyperbolic cosine function. The prediction model is iteratively trained based on the quantile loss function to obtain a trained probabilistic load prediction model. The trained probabilistic load forecasting model is used to carry out probabilistic load forecasting for urban and rural residents.

2. A probabilistic load forecasting method for urban and rural residents according to claim 1, characterized in that: The step of using the pre-processed historical load data as a sample set to train the prediction model based on the neural network includes: For the historical load data in the sample set, the time characteristics, meteorological characteristics and user characteristics in the historical load data are represented as numerical data in a one-hot encoding manner and then input into the prediction model.

3. The probabilistic load forecasting method for urban and rural residents according to claim 1 is characterized by: The prediction model adopts an LSTM neural network including an encoder and a decoder.

4. A probabilistic load forecasting method for urban and rural residents according to claim 1, characterized in that: The quantile loss function is specifically: Among them, q is the regression loss, which ranges from 0 to 1; y t is the true value, is the predicted value at the quantile at time t, and a is the parameter for adjusting the smoothness.

5. A probabilistic load forecasting system for urban and rural residents, characterized in that: include: A data acquisition module, used to acquire historical load data of urban and rural residents and perform preprocessing; A model training module is used to train a prediction model based on a neural network using preprocessed historical load data as a sample set. The prediction model uses a probabilistic load interval as a model output, and constructs a quantile loss function through a hyperbolic cosine function. The prediction model is iteratively trained based on the quantile loss function to obtain a trained probabilistic load prediction model. The probabilistic load forecasting module uses the trained probabilistic load forecasting model to perform probabilistic load forecasting for urban and rural residents.

6. A probabilistic load forecasting system for urban and rural residents according to claim 5, characterized in that: The step of using the pre-processed historical load data as a sample set to train the prediction model based on the neural network includes: For the historical load data in the sample set, the time characteristics, meteorological characteristics and user characteristics in the historical load data are represented as numerical data in a one-hot encoding manner and then input into the prediction model.

7. The probabilistic load forecasting system for urban and rural residents according to claim 5 is characterized by: The prediction model adopts an LSTM neural network including an encoder and a decoder.

8. A probabilistic load forecasting system for urban and rural residents according to claim 5, characterized in that: The quantile loss function is specifically: Among them, y t is the true value, is the predicted value at the quantile at time t, and a is the parameter for adjusting the smoothness.

9. An electronic device 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 probabilistic load forecasting method for urban and rural residents is implemented as described in any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the probabilistic load forecasting method for urban and rural residents as described in any one of claims 1 to 4 is implemented.

Images