Regional electricity price prediction method, device and system, and storage medium

Abstract

The invention discloses a regional electricity price prediction method, device and system, and a storage medium. The method comprises the following steps: S1, constructing a regional electricity price prediction function; s2, extracting regional electricity price data features; and S3, according to the regional electricity price prediction function and the regional electricity price data features, regional electricity price prediction is carried out through an LSTM deep learning structure. By adopting the technical scheme of the invention, the accuracy and real-time performance of electricity price prediction can be remarkably improved, a scientific basis and decision support are provided for participants of the electricity market, and the healthy development and sustainable development goal of the electricity market is promoted to be realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of power prediction, and particularly relates to a regional electricity price prediction method, device, system, and storage medium. Background Art

[0002] In the electricity market environment, electricity price is the core content of the entire market, and all parties participating in the market urgently need accurate electricity price prediction methods. For power generators, if they can accurately predict the market clearing price of the next day, it will help them formulate the optimal bidding strategy to obtain maximum profit. From the perspective of electricity purchasers, the electricity price constitutes their unit electricity purchase cost, and the prediction of electricity price makes it possible to control their dynamic costs; from the perspective of market regulators, the prediction of electricity price can provide a scientific basis for the healthy, stable, and orderly competition and development of the market and the formulation of various electricity price policies. Therefore, regional electricity price prediction is of great significance to all participants in the electricity market and has become an important part of the electricity market.

[0003] The accurate prediction of regional electricity prices and the formulation of efficient trading strategies not only concern the stable operation of the electricity market but also directly affect the economic benefits of energy enterprises and the electricity consumption costs of users. Traditional electricity price prediction mostly relies on statistical models or time series analysis methods. Although these methods can capture the electricity price fluctuation trend to a certain extent, when faced with the increasing volatility, nonlinearity, and high-dimensional data characteristics of the electricity market, their prediction accuracy and adaptability are insufficient. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a regional electricity price prediction method, device, system, and storage medium.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A regional electricity price prediction method includes the following steps:

[0007] Step S1: Construct a regional electricity price prediction function;

[0008] Step S2: Extract regional electricity price data features;

[0009] Step S3: According to the regional electricity price prediction function and regional electricity price data features, perform regional electricity price prediction through an LSTM deep learning structure.

[0010] Preferably, step S1 includes:

[0011] Use the least squares method to establish a regression function;

[0012] According to the regression function, construct a regional electricity price prediction function.

[0013] Preferably, step S2 includes:

[0014] Applying the grey relational analysis method to calculate the data feature correlation degree; wherein, the data features include: historical electricity price, power generation, climate conditions, and market factors;

[0015] Extracting the regional electricity price data features according to the level of the correlation degree.

[0016] The present invention also provides a regional electricity price prediction device, including:

[0017] A construction module for constructing a regional electricity price prediction function;

[0018] An extraction module for extracting the regional electricity price data features;

[0019] A prediction module for predicting the regional electricity price through an LSTM deep learning structure according to the regional electricity price prediction function and the regional electricity price data features.

[0020] Preferably, the construction module includes:

[0021] A first construction unit for establishing a regression function using the least squares method;

[0022] A second construction unit for constructing a regional electricity price prediction function according to the regression function.

[0023] Preferably, the extraction module includes:

[0024] A first extraction unit for applying the grey relational analysis method to calculate the data feature correlation degree; wherein, the data features include: historical electricity price, power generation, climate conditions, and market factors;

[0025] A second extraction unit for extracting the regional electricity price data features according to the level of the correlation degree.

[0026] The present invention also provides a regional electricity price prediction system, including: a memory and a processor, wherein a computer program is stored on the memory and run by the processor, and the computer program executes the regional electricity price prediction method when being run by the processor.

[0027] The present invention also provides a storage medium, on which a computer program is stored, and the computer program executes the regional electricity price prediction method when running.

[0028] The present invention predicts the regional electricity price by establishing a regional electricity price prediction function, extracting the characteristics of regional electricity price data, and performing regional electricity price prediction based on deep learning methods. With its powerful feature extraction and complex pattern recognition capabilities, deep learning can deeply explore the hidden laws in the electricity market, effectively address the non-linearity and high-noise problems in electricity price data, thereby significantly improving the accuracy and real-time performance of electricity price prediction, and promoting the development of the electricity market towards a more intelligent and efficient direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on the provided drawings without creative efforts.

[0030] Figure 1 It is a flowchart of the regional electricity price prediction method according to the embodiment of the present invention;

[0031] Figure 2 It is a graph of the electricity price fluctuation result of the embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0032] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0033] To make the above objects, features, and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the drawings and specific embodiments.

[0034] Embodiment 1:

[0035] As Figure 1 shown, the embodiment of the present invention provides a regional electricity price prediction method, including the following steps:

[0036] Step S1, construct a regional electricity price prediction function;

[0037] Step S2, extract the characteristics of regional electricity price data;

[0038] Step S3, according to the regional electricity price prediction function and the characteristics of regional electricity price data, perform regional electricity price prediction through the LSTM deep learning structure.

[0039] As an implementation manner of the embodiment of the present invention, step S1 includes:

[0040] Establish a regression function using the least squares method;

[0041] Construct a regional electricity price prediction function according to the regression function.

[0042] Furthermore, the electricity price in the electricity market has the special characteristic of non-linear change. Therefore, the least squares method is selected to establish a support vector regression function, and the function expression is as follows:

[0043]

[0044] ω represents the weight vector of the regression function; b represents the offset of the regression function. Take the electricity price data of a certain electricity market in the past two years as the training set, and set this training set as {(x i ,y i )}, i = 1, 2,..., n, where n represents the total number of training set samples, x i represents the i-th data index of the electricity price in the electricity market, and y i represents the i-th expected value of the electricity price in the electricity market. Project the training set into the high-dimensional feature space through the regression function and according to the principle of minimizing risk, the following loss function can be obtained:

[0045]

[0046] represents the expected error of the electricity price in the electricity market.

[0047] Furthermore, by analyzing the regression function and the loss function, the regional electricity price data can be divided into two different types, and the classification is as follows:

[0048]

[0049] Optimize the classification effect to better distinguish the regional electricity price data. The formula is as follows:

[0050]

[0051] Among them, K represents the output value of the regression model.

[0052] Through the above steps, the parameter value of the regional electricity price prediction function b can be obtained, so as to establish a decision function for regional electricity price prediction. The function expression is as follows:

[0053]

[0054] This prediction function simplifies the electricity price prediction solution process and improves the calculation efficiency and accuracy.

[0055] As an implementation manner of the embodiment of the present invention, step S2 includes:

[0056] Calculate the correlation degree of data features using the grey relational analysis method; among them, the data features include: historical electricity price, power generation, climate conditions, and market factors.

[0057] Extract the data features of regional electricity price according to the level of correlation degree.

[0058] Furthermore, in the prediction of regional electricity price in the power market, features such as historical electricity price, power generation, climate conditions, and market factors will all affect the prediction accuracy of regional electricity price. Therefore, in the process of feature extraction, these non-traditional data need to be converted into features that can be calculated.

[0059] The present invention calculates the correlation degree of data features using the grey relational analysis method. The higher the correlation degree, the closer the data features are. The calculation process is as follows.

[0060] First, define the input sample set as matrix D as follows:

[0061]

[0062] where λ n (k) represents the nth feature of the kth sample in the sample set.

[0063] According to matrix D, the expression of the target sequence can be calculated:

[0064] λ0 = [λ0(1), λ0(2), …, λ0(m)] T

[0065] Substitute the data features after standardization into the above formula to calculate the grey correlation degree between matrix λ and target sequence λ0. The formula is as follows:

[0066]

[0067] where Δ min represents the minimum value of the differences of each column in the matrix; Δ max represents the maximum value of the differences of each column in the matrix; ξ represents the correlation degree resolution coefficient; Δ 0i (t) represents the absolute value of the difference of the target sequence.

[0068] Finally, the grey correlation degree expression can be obtained as follows:

[0069]

[0070] According to the grey correlation degree of the above calculation results, the data features can be screened and extracted.

[0071] As an implementation manner of an embodiment of the present invention, in step S3, in the LSTM structure, in order to control the memory unit, the concepts of an input gate, a forget gate, and an output gate are introduced based on the regional electricity price prediction function. The specific architecture formula is as follows:

[0072] i t = σ(W i x t + U i h t-1 + b i )

[0073] f t = σ(W f x t + U f h t-1 + b f )

[0074] O t = σ(W o x t + U o h t-1 + b o )

[0075] W i 、W f 、W o respectively represent the weight matrices of the three gates; x t represents the input regional electricity price data features; h t-1 represents the state of the previous hidden unit; U i 、U f 、U o respectively represent the weight matrices connecting the state of the previous hidden unit; b i 、b f 、b o respectively represent the bias vectors of the three gates.

[0076] According to the LSTM structure, the regional electricity price is predicted based on the extracted historical electricity price data features. The output of each prediction result utilizes the available historical information data in the forward direction and combines the available future information data in the backward direction. This design can effectively avoid the problem of delayed output and can also avoid dispersing the electricity price data into different networks.

[0077] In the embodiment of the present invention, a regional electricity price prediction function is first established, then the characteristics of regional electricity price data are extracted, and finally, the regional electricity price is predicted based on the deep learning method. The least squares method is used to establish a regression function and the final regional electricity price prediction function is constructed on this basis; then the grey relational analysis method is applied to calculate the data feature correlation degree and the characteristics of regional electricity price data are extracted according to the level of the correlation degree; finally, the prediction of the regional electricity price is carried out based on the LSTM structure in the deep learning method. The present invention can significantly improve the accuracy and real-time performance of electricity price prediction, provide a scientific basis and decision-making support for the participants in the electricity market, and promote the healthy development of the electricity market and the realization of the sustainable development goal.

[0078] In order to verify the regional electricity price prediction effect of the present invention, in this embodiment, 55,000 data of the electricity price characteristics at different time periods of a certain day in a certain electricity market are used for simulation experiments on the built MATLAB experimental platform. The above data are divided into three groups: a training set, a validation set and a test set, and the prediction method is used to predict the electricity price fluctuation in the test set. The specific parameters of the three data sets are shown in Table 1.

[0079] Table 1

[0080]

[0081] The schematic diagram of the electricity price fluctuation drawn according to the prediction result is as Figure 2 shown. It can be seen from the prediction result that from 5 pm to 8 pm is the peak electricity consumption period, and the electricity price in this period reaches up to 0.65 yuan. Based on the prediction result of this method, a more reasonable design of the trading strategy for the regional electricity price can be carried out.

[0082] Embodiment 2:

[0083] The embodiment of the present invention also provides a regional electricity price prediction device, including:

[0084] A construction module for constructing a regional electricity price prediction function;

[0085] An extraction module for extracting the characteristics of regional electricity price data;

[0086] A prediction module for predicting the regional electricity price through the LSTM deep learning structure according to the regional electricity price prediction function and the characteristics of the regional electricity price data.

[0087] As an implementation manner of the embodiment of the present invention, the construction module includes:

[0088] A first construction unit for establishing a regression function using the least squares method;

[0089] A second construction unit for constructing a regional electricity price prediction function according to the regression function.

[0090] As an implementation manner of an embodiment of the present invention, the extraction module includes:

[0091] A first extraction unit, configured to calculate the data feature correlation degree by applying the grey relational analysis method; wherein, the data features include: historical electricity price, power generation, climate conditions, and market factors;

[0092] A second extraction unit, configured to extract the regional electricity price data features according to the level of the correlation degree.

[0093] Embodiment 3:

[0094] The embodiment of the present invention further provides a regional electricity price prediction system, including: a memory and a processor, where a computer program run by the processor is stored on the memory, and the computer program executes the regional electricity price prediction method when being run by the processor.

[0095] Embodiment 4:

[0096] The embodiment of the present invention further provides a storage medium, where a computer program is stored on the storage medium, and the computer program executes the regional electricity price prediction method when running.

[0097] The above embodiments are only descriptions of the preferred manners of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solution of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A regional electricity price forecasting method, characterized in that, It includes the following steps: Step S1, construct a regional electricity price prediction function; Step S2, extract the characteristics of regional electricity price data; Step S3, based on the regional electricity price prediction function and the characteristics of regional electricity price data, perform regional electricity price prediction through the LSTM deep learning structure.

2. The regional electricity price prediction method according to claim 1, characterized in that Step S1 includes: Use the least squares method to establish a regression function; Based on the regression function, construct a regional electricity price prediction function.

3. The regional electricity price prediction method according to claim 2, characterized in that, Step S2 includes: Apply the grey relational analysis method to calculate the data feature correlation degree; among them, the data features include: historical electricity price, power generation, climate conditions, market factors; Extract the characteristics of regional electricity price data according to the level of correlation degree.

4. A regional electricity price prediction device, characterized in that It includes: A construction module for constructing a regional electricity price prediction function; An extraction module for extracting the characteristics of regional electricity price data; A prediction module for performing regional electricity price prediction through the LSTM deep learning structure based on the regional electricity price prediction function and the characteristics of regional electricity price data.

5. The regional electricity price prediction device according to claim 4, wherein, The construction module includes: A first construction unit for using the least squares method to establish a regression function; A second construction unit for constructing a regional electricity price prediction function based on the regression function.

6. The regional electricity price prediction device according to claim 5, characterized in that, The extraction module includes: A first extraction unit for applying the grey relational analysis method to calculate the data feature correlation degree; among them, the data features include: historical electricity price, power generation, climate conditions, market factors; A second extraction unit for extracting the characteristics of regional electricity price data according to the level of correlation degree.

7. A regional electricity price forecasting system, characterized in that, It includes: A memory and a processor, where a computer program is stored on the memory and run by the processor, and the computer program, when run by the processor, executes the regional electricity price prediction method according to any one of claims 1-3.

8. A storage medium, characterized in that, A computer program is stored on the storage medium, and the computer program, when running, executes the regional electricity price prediction method according to any one of claims 1-3.

Images