CN116976927A - Short-term electricity price prediction method, system, computer and storage medium based on deep learning - Google Patents

Short-term electricity price prediction method, system, computer and storage medium based on deep learning Download PDF

Info

Publication number
CN116976927A
CN116976927A CN202310761691.6A CN202310761691A CN116976927A CN 116976927 A CN116976927 A CN 116976927A CN 202310761691 A CN202310761691 A CN 202310761691A CN 116976927 A CN116976927 A CN 116976927A
Authority
CN
China
Prior art keywords
data
electricity price
short
term
processing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310761691.6A
Other languages
Chinese (zh)
Inventor
王琳翰
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin Institute of Technology
Original Assignee
Harbin Institute of Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harbin Institute of Technology filed Critical Harbin Institute of Technology
Priority to CN202310761691.6A priority Critical patent/CN116976927A/en
Publication of CN116976927A publication Critical patent/CN116976927A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0201—Market modelling; Market analysis; Collecting market data
    • G06Q30/0206—Price or cost determination based on market factors
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/10—Pre-processing; Data cleansing
    • G06F18/15—Statistical pre-processing, e.g. techniques for normalisation or restoring missing data
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/02—Marketing; Price estimation or determination; Fundraising
    • G06Q30/0283—Price estimation or determination
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply

Abstract

A short-term electricity price prediction method, a short-term electricity price prediction system, a short-term electricity price prediction computer and a short-term electricity price prediction storage medium based on deep learning relate to the field of civil electricity price prediction. The problem that the existing short-term civil electricity price prediction model has strong volatility and randomness, and the price cannot be accurately predicted from the fluctuation of the electricity price in a unilateral manner is solved. The method comprises the following steps: acquiring electricity price data, humidity data, power load data and electricity consumption date, and carrying out Leida processing to acquire data after abnormality removal; acquiring linear relation among the abnormal electricity price data, the abnormal humidity data and the abnormal power load data; normalizing and wavelet transforming the data after the abnormality is removed; constructing an LSTM neural network model, and obtaining an electricity price prediction result; and calculating the electricity price according to the linear relation among the abnormal electricity price data, the humidity data and the power load data, and carrying out weighted average on the electricity price and the predicted result to obtain a short-term civil electricity price predicted result. The method is applied to the reserved charging field of the electric automobile.

Description

Short-term electricity price prediction method, system, computer and storage medium based on deep learning
Technical Field
The invention relates to the field of prediction of civil electricity prices, in particular to a short-term electricity price prediction method based on deep learning.
Background
With the continuous increase of electric vehicles, the charging problem of the electric vehicles gradually becomes a social problem, and the corresponding electricity price becomes a main problem considered by the owners of the electric vehicles in charging. If a user can predict the electricity price at a certain moment in advance, a charging plan is made, and reserved charging is carried out in an optimal time period, so that the charging cost is greatly reduced. The reserved charging service derived through electricity price prediction can also enable the charging station to acquire the passenger flow load condition of the charging station in advance, and power dispatching and adjustment are performed in advance, so that the safety of a power grid is protected. The research of the electricity price prediction method has great value for both the benefits of charging users and the power dispatching between charging stations.
The method widely applied to the current power market spot transaction electricity price prediction comprises the following steps: the time sequence model, the artificial neural network prediction method and the traditional similar daily method are used for predicting the electricity price, but the electricity price is influenced by various factors such as supply and demand relation, climate factors, power supply paths, power loads and the like, has strong volatility and randomness, one side cannot accurately predict the price from the fluctuation of the electricity price, and needs to consider influence factors in various aspects as much as possible to find the potential influence degree of each factor on the electricity price.
Disclosure of Invention
The invention provides a short-term electricity price prediction method based on deep learning, which aims at solving the problem that the existing short-term civil electricity price prediction model has strong volatility and randomness, and unilateral price cannot be accurately predicted from the fluctuation of electricity price.
The short-term electricity price prediction method based on deep learning is realized by the following technical scheme that:
acquiring electricity price data, humidity data, power load data and electricity consumption date;
carrying out Laida processing on the electricity price data, the humidity data and the power load data to obtain data after abnormality removal;
acquiring linear relations among the power price data, the humidity data and the power load data after the abnormality is removed according to the data after the abnormality is removed;
normalizing the data after the abnormality is removed;
performing wavelet transformation processing on the normalized data;
constructing an LSTM neural network model according to the data after wavelet processing;
acquiring an electricity price prediction result according to the LSTM neural network model;
and calculating the electricity price according to the linear relation among the abnormal electricity price data, the humidity data and the power load data, and carrying out weighted average on the electricity price and the predicted result to obtain a short-term civil electricity price predicted result.
Further, there is provided a preferable mode, wherein the step of performing rayleigh processing on the electricity price data, the humidity data and the power load data to obtain data after abnormality removal includes:
calculating variances and mean values of the electricity price data, the humidity data and the power load data;
determining data of variance three times different from the mean value as abnormal data;
and replacing the abnormal data with the mean value data to finish the Leida processing.
Further, a preferred mode is provided, wherein the normalizing processing is performed on the data after the abnormality is removed, specifically:
where d is normalized data, d i D, as the current data before normalization min D is the minimum value in all data max Is the maximum of all data.
Further, there is provided a preferred mode, wherein the wavelet transformation processing is performed on the normalized data, specifically:
where WT (a, τ) is the result of wavelet transform, a is the scale factor, τ is the amount of translation, f (t) is the signal to be analyzed, and t is the time domain signal.
Further, a preferred mode is also provided, and an LSTM neural network model is constructed according to the data after wavelet processing, specifically:
the LSTM neural network model comprises an input layer, an LSTM hidden layer, a full connection layer and an output layer;
the input layer is used for receiving normalized data and outputting the normalized data to the LSTM hidden layer for training and prediction;
the hidden layer is connected with the input layer and is used for receiving data and predicting output data at the next moment, and the output data at the next moment is transmitted to the full-connection layer;
the full-connection layer is connected with the hidden layer and is used for receiving data output by the hidden layer, and processing and dimension reducing the characteristics extracted by the hidden layer;
the output layer is connected with the full-connection layer and is used for carrying out weighted sum processing on the output of the full-connection layer and finally outputting a prediction result.
Based on the same inventive concept, the invention also provides a short-term electricity price prediction system based on deep learning, which comprises:
the data acquisition unit is used for acquiring electricity price data, humidity data, power load data and electricity utilization date;
the abnormal data processing unit is used for carrying out the Leida processing on the electricity price data, the humidity data and the electric load data to obtain data after the abnormality is removed;
a linear relation obtaining unit, configured to obtain a linear relation of the power price data, the humidity data and the power load data from which the abnormality is removed according to the data from which the abnormality is removed;
the normalization unit is used for performing normalization processing on the data after the abnormality is removed;
a wavelet transformation processing unit for performing wavelet transformation processing on the normalized data;
the LSTM neural network model building unit is used for building an LSTM neural network model according to the data after the wavelet processing;
the electricity price prediction unit is used for obtaining an electricity price prediction result according to the LSTM neural network model;
and the short-term civil electricity price prediction result acquisition unit is used for calculating electricity price according to the linear relation among the abnormality-removed electricity price data, the humidity data and the power load data, and carrying out weighted average on the electricity price and the prediction result to acquire a short-term civil electricity price prediction result.
Further, there is also provided a preferable mode, the abnormal data processing unit including:
calculating variances and mean values of the electricity price data, the humidity data and the power load data;
determining data of variance three times different from the mean value as abnormal data;
and replacing the abnormal data with the mean value data to finish the Leida processing.
Further, there is also provided a preferred mode, the normalization unit specifically includes:
where d is normalized data, d i D, as the current data before normalization min D is the minimum value in all data max Is the maximum of all data.
Based on the same inventive concept, the present invention also provides a computer-readable storage medium for storing a computer program that performs a short-term electricity price prediction method based on deep learning as described in any one of the above.
Based on the same inventive concept, the present invention also provides a computer device comprising a memory and a processor, the memory storing a computer program, the processor executing a short-term electricity price prediction method based on deep learning according to any one of the above, when the processor runs the computer program stored in the memory.
The invention has the advantages that:
the invention solves the problems that the existing short-term civil electricity price prediction model has strong volatility and randomness, and the price cannot be accurately predicted from the fluctuation of the electricity price.
The short-term electricity price prediction method based on deep learning is different from the traditional prediction method which only considers the electricity price floating rule to predict the electricity price, and the method takes humidity and electric load which influence the electricity price as reference factors influencing the electricity price, and also utilizes the characteristics of LSTM neural network which are good in finding out the regularity and the trend of time sequence data to predict the same, and obtains corresponding electricity price prediction results according to the linear relation existing between the LSTM neural network and the LSTM neural network. The invention uses wavelet transformation to preprocess non-stationary random data, eliminates the influence of noise on prediction precision, optimizes data set, enhances the anti-interference capability of model, and improves the accuracy of prediction.
In addition, the invention realizes analysis of electricity price from multiple aspects by adding consideration of humidity and electric load, optimizes the existing prediction method and improves the prediction accuracy.
The method is applied to the reserved charging field of the electric automobile.
Drawings
Fig. 1 is a flowchart of a short-term electricity price prediction method based on deep learning according to an embodiment;
FIG. 2 is a hierarchical diagram of an LSTM neural network model according to a fifth embodiment;
fig. 3 is a schematic diagram of a predicted electricity price according to an eleventh embodiment;
fig. 4 is a schematic diagram showing the result of prediction of electricity prices in half a month by the prediction model according to the eleventh embodiment.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments.
Embodiment one, this embodiment will be described with reference to fig. 1. The embodiment relates to a short-term electricity price prediction method based on deep learning, which comprises the following steps of;
acquiring electricity price data, humidity data, power load data and electricity consumption date;
carrying out Laida processing on the electricity price data, the humidity data and the power load data to obtain data after abnormality removal;
acquiring linear relations among the power price data, the humidity data and the power load data after the abnormality is removed according to the data after the abnormality is removed;
normalizing the data after the abnormality is removed;
performing wavelet transformation processing on the normalized data;
constructing an LSTM neural network model according to the data after wavelet processing;
acquiring an electricity price prediction result according to the LSTM neural network model;
and calculating the electricity price according to the linear relation among the abnormal electricity price data, the humidity data and the power load data, and carrying out weighted average on the electricity price and the predicted result to obtain a short-term civil electricity price predicted result.
According to the short-term electricity price prediction method based on deep learning, various data such as electricity price data, humidity data and power load data are comprehensively used, information of different dimensions is fully utilized to model electricity price prediction, and accuracy and stability of prediction can be improved; abnormal data can be effectively processed through the Leida processing and the normalization processing, and the data range is scaled to be within the standard range, so that the influence of data abnormal points on the model is reduced, and the robustness of the model is improved; the wavelet analysis method is adopted to preprocess the data, so that the local characteristics and the change trend of the data can be better found, the abnormal condition in the data can be better detected and processed, and the analysis precision and the prediction effect of the data are improved; by constructing a neural network model, time sequence data can be effectively processed, and the change of electricity price time sequence can be captured and predicted, so that the neural network model has strong expression capability and prediction accuracy; the linear relation is utilized to calculate the electricity price, so that the relation of the electricity price influenced by various factors can be better explored, and the accuracy and stability of electricity price prediction are improved.
According to the short-term electricity price prediction method based on deep learning, a comprehensive plurality of data are provided for the problem of electricity price prediction, the influence of various data on electricity price trend is fully utilized, prediction is carried out through a neural network model, and appropriate correction is carried out through weighting of other factors, so that the accuracy and practical value of a prediction result are improved.
In a second embodiment, the present embodiment is a further limitation of the short-term electricity price prediction method based on deep learning according to the first embodiment, wherein the performing the rayleigh process on the electricity price data, the humidity data, and the power load data to obtain the data after the abnormality is removed includes:
calculating variances and mean values of the electricity price data, the humidity data and the power load data;
determining data of variance three times different from the mean value as abnormal data;
and replacing the abnormal data with the mean value data to finish the Leida processing.
The rada of the present embodiment processes the electricity price data, the humidity data, and the power load data, and mainly aims to detect and replace abnormal values so as to improve the quality, accuracy, and reliability of the data. When the power price data, the humidity data and the electric load data are subjected to the Leida processing, the variance and the mean value of the data are calculated to determine which data points in the data are far away from the mean value, namely the abnormal value. Replacing the outlier data that differs from the mean by three times the variance with the mean may achieve several objectives:
1. removing abnormal data: abnormal data may have adverse effects on the distribution and statistical analysis of the data, which may be removed by the raydad process to better understand and analyze the data.
2. Normal data is retained: the Leida processing only replaces abnormal data far away from the average value, other normal data are reserved, and the influence of data disturbance on the overall result is reduced.
3. Stability of data is improved: replacing the anomalous data with the average value may reduce fluctuations and uncertainty in the data, making the data smoother and more stable.
In a third embodiment, the present embodiment is a further limitation of the short-term electricity price prediction method based on deep learning according to the first embodiment, wherein the normalizing processing is performed on the data after the abnormality removal, specifically:
where d is normalized data, d i D, as the current data before normalization min D is the minimum value in all data max Is the maximum of all data.
In the present embodiment, the data from which the abnormality has been removed is normalized, and the obtained data is new data scaled to a certain range. The normalization process can reduce the absolute value range of the data to be within a standard range, and is beneficial to comparison and analysis between different data. The present embodiment uses a maximum-minimum normalization process to map the data into the [0,1] range. Through normalization processing, the relation and trend among the data can be better explored, so that the use efficiency and analysis accuracy of the data are improved.
In a fourth embodiment, the present embodiment is a further limitation of the short-term electricity price prediction method according to the first embodiment, wherein the performing of the wavelet transform processing on the normalized data specifically includes:
where WT (a, τ) is the result of wavelet transform, a is the scale factor, τ is the amount of translation, f (t) is the signal to be analyzed, and t is the time domain signal.
In the embodiment, wavelet transformation processing is introduced, and because a large amount of data is needed for electricity price prediction through the LSTM neural network model, but a large amount of data is often accompanied with the problem of noise, the introduction of wavelet transformation can realize the conversion of data from a time domain to a frequency domain, eliminate the influence of noise on the data, observe the essential characteristics and fluctuation rule of the data in the frequency domain, and improve the accuracy of LSTM model prediction.
Embodiment five, this embodiment will be described with reference to fig. 2. The present embodiment is further defined by the short-term electricity price prediction method based on deep learning according to the first embodiment, wherein the constructing an LSTM neural network model according to the data after wavelet processing specifically includes:
the LSTM neural network model comprises an input layer, an LSTM hidden layer, a full connection layer and an output layer;
the input layer is used for receiving normalized data and outputting the normalized data to the LSTM hidden layer for training and prediction;
the hidden layer is connected with the input layer and is used for receiving data and predicting output data at the next moment, and the output data at the next moment is transmitted to the full-connection layer;
the full-connection layer is connected with the hidden layer and is used for receiving data output by the hidden layer, and processing and dimension reducing the characteristics extracted by the hidden layer;
the output layer is connected with the full-connection layer and is used for carrying out weighted sum processing on the output of the full-connection layer and finally outputting a prediction result.
The main structure of the LSTM neural network model constructed in the embodiment comprises an input layer, an LSTM hidden layer, a full connection layer and an output layer.
Input layer: the data is used for receiving normalized data and providing the normalized data for an LSTM hidden layer for training and prediction.
LSTM hidden layer: in the LSTM hidden layer, the LSTM model is adopted to process time series data, and history information for learning data comprises the LSTM unit state at the last moment, the output at the last moment, the input at the current moment and the like, so that the output at the next moment is predicted. Since the LSTM hidden layer has a strong memory capacity and long-term dependency, time-series data can be processed well.
Full tie layer: and one or more full-connection layers are added behind the LSTM hidden layer and used for processing and reducing the dimension of the features extracted by the LSTM hidden layer, so that the performance and generalization capability of the model are improved, and the risk of overfitting is reduced. The fully connected layer may be designed using different activation functions and regularization methods to achieve optimal performance and prediction results.
Output layer: and the output layer carries out weighted sum processing on the output of the full-connection layer, and finally outputs a prediction result. In actual modeling, the output layer can be designed according to the needs, including output dimensions, activation functions and the like.
A sixth embodiment is a short-term electricity price prediction system based on deep learning according to the present embodiment, including:
the data acquisition unit is used for acquiring electricity price data, humidity data, power load data and electricity utilization date;
the abnormal data processing unit is used for carrying out the Leida processing on the electricity price data, the humidity data and the electric load data to obtain data after the abnormality is removed;
a linear relation obtaining unit, configured to obtain a linear relation of the power price data, the humidity data and the power load data from which the abnormality is removed according to the data from which the abnormality is removed;
the normalization unit is used for performing normalization processing on the data after the abnormality is removed;
a wavelet transformation processing unit for performing wavelet transformation processing on the normalized data;
the LSTM neural network model building unit is used for building an LSTM neural network model according to the data after the wavelet processing;
the electricity price prediction unit is used for obtaining an electricity price prediction result according to the LSTM neural network model;
and the short-term civil electricity price prediction result acquisition unit is used for calculating electricity price according to the linear relation among the abnormality-removed electricity price data, the humidity data and the power load data, and carrying out weighted average on the electricity price and the prediction result to acquire a short-term civil electricity price prediction result.
An embodiment seven, this embodiment is a further limitation of the short-term electricity price prediction system based on deep learning according to the sixth embodiment, wherein the anomaly data processing unit includes:
calculating variances and mean values of the electricity price data, the humidity data and the power load data;
determining data of variance three times different from the mean value as abnormal data;
and replacing the abnormal data with the mean value data to finish the Leida processing.
An eighth embodiment is a further limitation of the short-term electricity price prediction system according to the sixth embodiment, wherein the normalization unit specifically includes:
where d is normalized data, d i D, as the current data before normalization min D is the minimum value in all data max Is the maximum of all data.
The computer-readable storage medium according to the ninth embodiment is a computer program for executing the short-term electricity price prediction method according to any one of the first to fifth embodiments.
A computer device according to a tenth embodiment includes a memory in which a computer program is stored and a processor that executes a short-term electricity price prediction method based on deep learning according to any one of the first to fifth embodiments when the processor runs the computer program stored in the memory.
Embodiment eleven, this embodiment will be described with reference to fig. 3 and 4. The present embodiment provides a specific example for the short-term electricity price prediction method based on deep learning in the first embodiment, and is also used for explaining the second embodiment to the fifth embodiment, specifically:
step one: acquiring historical power price, power load, date and other relevant data disclosed in the disclosure of the power price from a power price transaction center, extracting power price once every 30 minutes, and extracting data of 48 power prices in one day and 5 years in 2016 to 2020;
step two: carrying out Laida processing on electricity price data, humidity data and electric load data, respectively calculating variance and mean value of each data, and replacing abnormal data which are different from the mean value by three times of variance with the mean value;
step three: normalizing the data
Where d is normalized data, d i D, as the current data before normalization min D is the minimum value in all data max Maximum value among all data;
step four: and carrying out wavelet transformation processing on the normalized data, wherein a wavelet transformation formula is as follows:
step five: constructing an LSTM neural network model;
step six: training the LSTM model by utilizing the historical electricity price data, the historical humidity data and the historical power load data respectively;
step seven: respectively obtaining each prediction result;
step eight: the linear relation between electricity price, humidity and electric load is found, a correlation coefficient is calculated, and the strength of the correlation is judged;
step nine: and calculating the electricity price through the linear relation between the humidity and the power load and the electricity price, and obtaining a final prediction result through weighted average of the electricity price and the prediction.
Step ten: and testing the prediction result by using the historical data, and randomly extracting the accuracy of the prediction result by checking the electricity price for a plurality of days.
As shown in fig. 3, the predicted electricity price is obtained on the 12 th month 31 of 2020, wherein blue line True Data is real electricity price Data, and yellow line Prediction is obtained as a predicted result, and it can be seen that the predicted result has peak Data at some time but the overall trend is approximately the same as that of the real Data, and has strong reference.
As shown in fig. 4, the prediction model shows that the method provided by the embodiment has a good effect on short-term prediction, also has a good prediction result in medium-term and long-term prediction, has a prediction trend approximately the same as that of the real electricity price, has strong practical applicability, and has a certain guiding significance on production and life.
While the preferred embodiments of the present disclosure have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the disclosure. It will be apparent to those skilled in the art that various modifications and variations can be made to the present disclosure without departing from the spirit or scope of the disclosure. Thus, the present disclosure is intended to include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
It will be appreciated by those skilled in the art that embodiments of the present disclosure may be provided as a method, system, or computer program product. Accordingly, the present disclosure may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present disclosure may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
The present disclosure is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be noted that: the above embodiments are only for illustrating the technical aspects of the present disclosure and not for limiting the scope thereof, and although the present disclosure has been described in detail with reference to the above embodiments, it should be understood by those of ordinary skill in the art that: various alterations, modifications, and equivalents may be suggested to the specific embodiments of the invention, which would occur to persons skilled in the art upon reading the disclosure, are intended to be within the scope of the appended claims.

Claims (10)

1. A short-term electricity price prediction method based on deep learning, characterized in that the method comprises;
acquiring electricity price data, humidity data, power load data and electricity consumption date;
carrying out Laida processing on the electricity price data, the humidity data and the power load data to obtain data after abnormality removal;
acquiring linear relations among the power price data, the humidity data and the power load data after the abnormality is removed according to the data after the abnormality is removed;
normalizing the data after the abnormality is removed;
performing wavelet transformation processing on the normalized data;
constructing an LSTM neural network model according to the data after wavelet processing;
acquiring an electricity price prediction result according to the LSTM neural network model;
and calculating the electricity price according to the linear relation among the abnormal electricity price data, the humidity data and the power load data, and carrying out weighted average on the electricity price and the predicted result to obtain a short-term civil electricity price predicted result.
2. The short-term electricity price prediction method based on deep learning according to claim 1, wherein the performing the raydad process on the electricity price data, the humidity data and the power load data to obtain the data after the abnormality is removed comprises:
calculating variances and mean values of the electricity price data, the humidity data and the power load data;
determining data of variance three times different from the mean value as abnormal data;
and replacing the abnormal data with the mean value data to finish the Leida processing.
3. The short-term electricity price prediction method based on deep learning according to claim 1, wherein the normalizing processing is performed on the data after removing the abnormality, specifically:
where d is normalized data, d i D, as the current data before normalization min D is the minimum value in all data max Is the maximum of all data.
4. The short-term electricity price prediction method based on deep learning according to claim 1, wherein the wavelet transformation processing is performed on the normalized data, specifically:
where WT (a, τ) is the result of wavelet transform, a is the scale factor, τ is the amount of translation, f (t) is the signal to be analyzed, and t is the time domain signal.
5. The short-term electricity price prediction method based on deep learning according to claim 1, wherein the constructing of the LSTM neural network model according to the wavelet processed data is specifically as follows:
the LSTM neural network model comprises an input layer, an LSTM hidden layer, a full connection layer and an output layer;
the input layer is used for receiving normalized data and outputting the normalized data to the LSTM hidden layer for training and prediction;
the hidden layer is connected with the input layer and is used for receiving data and predicting output data at the next moment, and the output data at the next moment is transmitted to the full-connection layer;
the full-connection layer is connected with the hidden layer and is used for receiving data output by the hidden layer, and processing and dimension reducing the characteristics extracted by the hidden layer;
the output layer is connected with the full-connection layer and is used for carrying out weighted sum processing on the output of the full-connection layer and finally outputting a prediction result.
6. A short-term electricity price prediction system based on deep learning, the system comprising:
the data acquisition unit is used for acquiring electricity price data, humidity data, power load data and electricity utilization date;
the abnormal data processing unit is used for carrying out the Leida processing on the electricity price data, the humidity data and the electric load data to obtain data after the abnormality is removed;
a linear relation obtaining unit, configured to obtain a linear relation of the power price data, the humidity data and the power load data from which the abnormality is removed according to the data from which the abnormality is removed;
the normalization unit is used for performing normalization processing on the data after the abnormality is removed;
a wavelet transformation processing unit for performing wavelet transformation processing on the normalized data;
the LSTM neural network model building unit is used for building an LSTM neural network model according to the data after the wavelet processing;
the electricity price prediction unit is used for obtaining an electricity price prediction result according to the LSTM neural network model;
and the short-term civil electricity price prediction result acquisition unit is used for calculating electricity price according to the linear relation among the abnormality-removed electricity price data, the humidity data and the power load data, and carrying out weighted average on the electricity price and the prediction result to acquire a short-term civil electricity price prediction result.
7. A deep learning based short term electricity price prediction system according to claim 6, characterized in that the anomaly data processing unit comprises:
calculating variances and mean values of the electricity price data, the humidity data and the power load data;
determining data of variance three times different from the mean value as abnormal data;
and replacing the abnormal data with the mean value data to finish the Leida processing.
8. The short-term civil electricity price prediction system based on the LSTM neural network according to claim 6, wherein the normalization unit is specifically:
where d is normalized data, d i D, as the current data before normalization min D is the minimum value in all data max Is the maximum of all data.
9. A computer-readable storage medium, characterized in that the computer-readable storage medium is for storing a computer program that performs a short-term price prediction method based on deep learning according to any one of claims 1-5.
10. A computer device, characterized by: comprising a memory and a processor, the memory having stored therein a computer program, which when executed by the processor performs a short-term electricity price prediction method based on deep learning as claimed in any one of claims 1-5.
CN202310761691.6A 2023-06-26 2023-06-26 Short-term electricity price prediction method, system, computer and storage medium based on deep learning Pending CN116976927A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310761691.6A CN116976927A (en) 2023-06-26 2023-06-26 Short-term electricity price prediction method, system, computer and storage medium based on deep learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310761691.6A CN116976927A (en) 2023-06-26 2023-06-26 Short-term electricity price prediction method, system, computer and storage medium based on deep learning

Publications (1)

Publication Number Publication Date
CN116976927A true CN116976927A (en) 2023-10-31

Family

ID=88472058

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310761691.6A Pending CN116976927A (en) 2023-06-26 2023-06-26 Short-term electricity price prediction method, system, computer and storage medium based on deep learning

Country Status (1)

Country Link
CN (1) CN116976927A (en)

Similar Documents

Publication Publication Date Title
CN110610280B (en) Short-term prediction method, model, device and system for power load
CN113962364B (en) Multi-factor power load prediction method based on deep learning
Cheng et al. Enhanced state estimation and bad data identification in active power distribution networks using photovoltaic power forecasting
CN103093285A (en) Short-term load forecast method based on artificial neural network
CN111680841B (en) Short-term load prediction method, system and terminal equipment based on principal component analysis
CN113126019A (en) Intelligent ammeter error remote estimation method, system, terminal and storage medium
CN115496627A (en) Method and system for evaluating response potential of adjustable resource
CN115034485A (en) Wind power interval prediction method and device based on data space
CN111985695A (en) Urban complex electricity consumption prediction method and device, electronic equipment and storage medium
CN117172589A (en) Industry typical user power demand response potential evaluation method
CN110390440B (en) Clustering and deep neural network-based intelligent ammeter user aggregate load prediction method
CN112215398A (en) Power consumer load prediction model establishing method, device, equipment and storage medium
CN105488598A (en) Medium-and-long time electric power load prediction method based on fuzzy clustering
CN114266416A (en) Photovoltaic power generation power short-term prediction method and device based on similar days and storage medium
CN113869590A (en) EEMD-LSTM-based regional energy internet load prediction method and system
Wan et al. Neural network model comparison and analysis of prediction methods using ARIMA and LSTM models
CN112598181A (en) Load prediction method, device, equipment and storage medium
CN117113086A (en) Energy storage unit load prediction method, system, electronic equipment and medium
CN116976927A (en) Short-term electricity price prediction method, system, computer and storage medium based on deep learning
CN116861306A (en) Abnormal electricity utilization detection method based on electricity utilization trend quantification network
Majidpour Time series prediction for electric vehicle charging load and solar power generation in the context of smart grid
CN115166625A (en) Intelligent ammeter error estimation method and device
CN114140176A (en) Adjustable capacity prediction method and device for load aggregation platform
Tao et al. Load forecasting based on short-term correlation clustering
Deng et al. Medium-term rolling load forecasting based on seasonal decomposition and long short-term memory neural network

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination