CN115204467A - Power load prediction method, device and storage medium - Google Patents
Power load prediction method, device and storage medium Download PDFInfo
- Publication number
- CN115204467A CN115204467A CN202210710222.7A CN202210710222A CN115204467A CN 115204467 A CN115204467 A CN 115204467A CN 202210710222 A CN202210710222 A CN 202210710222A CN 115204467 A CN115204467 A CN 115204467A
- Authority
- CN
- China
- Prior art keywords
- data
- online
- power load
- time information
- load prediction
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 74
- 238000003860 storage Methods 0.000 title claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 29
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 25
- 238000007781 pre-processing Methods 0.000 claims abstract description 25
- 238000005516 engineering process Methods 0.000 claims abstract description 16
- 230000008569 process Effects 0.000 claims abstract description 12
- 238000010276 construction Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims description 32
- 238000013507 mapping Methods 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 12
- 238000006243 chemical reaction Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 11
- 238000011176 pooling Methods 0.000 claims description 9
- 125000004122 cyclic group Chemical group 0.000 claims description 7
- 230000010354 integration Effects 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 16
- 238000010586 diagram Methods 0.000 description 12
- 238000004458 analytical method Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000014509 gene expression Effects 0.000 description 2
- 238000012886 linear function Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000007704 transition Effects 0.000 description 2
- 230000003213 activating effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 230000002194 synthesizing effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
Images
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
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- 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
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Strategic Management (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- General Business, Economics & Management (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Marketing (AREA)
- Computational Linguistics (AREA)
- Pure & Applied Mathematics (AREA)
- Evolutionary Computation (AREA)
- Operations Research (AREA)
- Computational Mathematics (AREA)
- Molecular Biology (AREA)
- Tourism & Hospitality (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Quality & Reliability (AREA)
Abstract
The invention discloses a power load prediction method, a device and a storage medium, wherein the method comprises the following steps: acquiring meteorological data and corresponding time information on line; preprocessing the online acquired meteorological data and corresponding time information to generate online comprehensive data; converting the online comprehensive data into an online data image based on an image processing technology; inputting an online data image into a pre-constructed power load prediction model to obtain a power load prediction value; wherein, the construction process of the power load prediction model comprises the following steps: constructing a convolutional neural network combined with an attention module, and training offline data images and corresponding power load actual values respectively as the input and the output of the convolutional neural network to generate a power load prediction model; the method can synthesize meteorological and time information, and improve the prediction precision of the model and reduce the complexity through an attention mechanism.
Description
Technical Field
The invention relates to a power load prediction method, a power load prediction device and a storage medium, and belongs to the technical field of power systems.
Background
With the development of industry and agriculture and the increasing living standard of people, the demand of society for electric power is larger and larger. To meet the increasing demand for electric power, the size of electric power systems must be continuously enlarged. Because the development of the power industry needs huge investment and primary energy and has huge influence on other departments of national economy, the reasonable power system planning can obtain huge economic benefits and huge social benefits. In contrast, a misplanning of an electric power system may bring irreparable losses to national construction. Therefore, the analysis and research on the power planning to obtain the maximum improvement of the planning quality have great significance, and the first step of achieving the goal is to make good power load prediction. The accuracy of the power load prediction directly affects the rationality of investment, network layout and operation.
Various prediction methods are generated, and the commonly used short-term prediction method of the power load mainly comprises a time sequence method and a function prediction method. The time sequence method realizes the prediction of the load value in the future time interval by presetting the length of the time interval and combining the load related information in the period with the load value data for analysis and comparison, wherein the set time interval can be one month, one quarter or one year; the function prediction method is based on a statistical theory and combines the analysis of the power load characteristics and various function expressions to construct a prediction function in advance. The method needs a set of reasonable weight distribution calculation principle to set the influence capacity of various load characteristics on the load value, the weight distribution is used as the input of a prediction model, the prediction model is divided into units, and then functions and weight values in each unit are subdivided, but the main problem is that the method has high searching difficulty for reasonable model functions and weight settings, the method depends on the long-term prediction experience of an operator on load prediction, and when the model is added with artificial subjective selection, the accuracy of a prediction result is not high.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provide a power load prediction method, a power load prediction device and a storage medium, which integrate meteorological and time information and improve the prediction precision of a model and reduce the complexity through an attention mechanism.
In order to achieve the purpose, the invention adopts the following technical scheme:
in a first aspect, the present invention provides a power load prediction method, including:
acquiring meteorological data and corresponding time information on line;
preprocessing the online acquired meteorological data and corresponding time information to generate online comprehensive data;
converting the online comprehensive data into an online data image based on an image processing technology;
inputting an online data image into a pre-constructed power load prediction model to obtain a power load prediction value;
wherein, the construction process of the power load prediction model comprises the following steps:
acquiring meteorological data, corresponding time information and an actual value of a power load offline;
preprocessing the meteorological data acquired offline and corresponding time information to generate offline comprehensive data;
converting the offline comprehensive data into an offline data image based on an image processing technology;
and constructing a convolutional neural network combined with an attention module, and training the offline data image and the corresponding power load actual value respectively as the input and the output of the convolutional neural network to generate a power load prediction model.
Optionally, the preprocessing the online acquired meteorological data and the corresponding time information to generate online comprehensive data includes:
meteorological data: normalization was performed for each meteorological datum:
in the formula, x i,n 、x′ i,n Respectively the value, x, of the ith meteorological data before and after normalization i,max 、x i,min The maximum value and the minimum value of the ith meteorological data are obtained;
time information: twelve month portions in one year are coded according to the number of months, and twenty-four hours in one day are respectively coded into 1, 2, 3 and 4 according to 0-6, 6-12, 12-18 and 18-24;
and (3) online comprehensive data: merging the normalized meteorological data and the encoded time information:
x n ＝[x 1,n ,…,x m,n ,t n ,d n ]
in the formula, x n For the nth online integration data, x m,n Is the m-th meteorological data, t n And d n The number of months and hours respectively.
Optionally, the converting the online comprehensive data into the online data image based on the image processing technology includes:
converting the online comprehensive data into an online comprehensive data matrix by a cyclic shift method;
carrying out interpolation processing on the online comprehensive data matrix by a cubic spline interpolation method to complete the dimension expansion of the online comprehensive data matrix;
and mapping the element values in the online comprehensive data matrix after the dimensionality expansion into pixel values in an RGB image by a linear mapping method to form an online data image.
Optionally, the convolutional neural network includes a first convolutional layer, a pooling layer, an attention module, a second convolutional layer, a full-link layer, and an output layer, which are connected in sequence; the attention module includes a channel attention module and a spatial attention module combined in a serial manner.
In a second aspect, the present invention provides an electrical load prediction apparatus, the apparatus comprising:
the online data module is used for acquiring meteorological data and corresponding time information online;
the online preprocessing module is used for preprocessing the online acquired meteorological data and the corresponding time information to generate online comprehensive data;
the online conversion module is used for converting the online comprehensive data into an online data image based on an image processing technology;
the online prediction module is used for inputting online data images into a pre-constructed power load prediction model to obtain a power load prediction value;
wherein, the construction process of the power load prediction model comprises the following steps:
the off-line data module is used for off-line acquiring meteorological data, corresponding time information and an actual value of the power load;
the off-line preprocessing module is used for preprocessing the off-line acquired meteorological data and the corresponding time information to generate off-line comprehensive data;
the offline conversion module is used for converting the offline comprehensive data into an offline data image based on an image processing technology;
and the model building module is used for building a convolutional neural network combined with the attention module, and training the offline data image and the corresponding power load actual value as the input and the output of the convolutional neural network respectively to generate a power load prediction model.
Optionally, the preprocessing the online acquired meteorological data and the corresponding time information to generate online comprehensive data includes:
meteorological data: normalization was performed for each meteorological data:
in the formula, x i,n 、x′ i,n Respectively the value, x, of the ith meteorological data before and after normalization i,max 、x i,min The maximum value and the minimum value of the ith meteorological data are obtained;
time information: twelve months in a year are coded according to the number of months, and twenty-four hours in a day are respectively coded as 1, 2, 3 and 4 according to 0-6, 6-12, 12-18 and 18-24;
and (3) online comprehensive data: merging the normalized meteorological data and the encoded time information:
x n ＝[x 1,n ,…,x m,n ,t n ,d n ]
in the formula, x n For the nth online integration data, x m,n Is the m-th meteorological data, t n And d n The number of months and hours respectively.
Optionally, the converting the online comprehensive data into the online data image based on the image processing technology includes:
converting the online comprehensive data into an online comprehensive data matrix by a cyclic shift method;
carrying out interpolation processing on the online comprehensive data matrix by a cubic spline interpolation method to complete the dimension expansion of the online comprehensive data matrix;
and mapping the element values in the online comprehensive data matrix after the dimensionality expansion into pixel values in an RGB image by a linear mapping method to form an online data image.
Optionally, the convolutional neural network includes a first convolutional layer, a pooling layer, an attention module, a second convolutional layer, a full-link layer, and an output layer, which are connected in sequence; the attention module includes a channel attention module and a spatial attention module combined in a serial manner.
In a third aspect, the present invention provides an electrical load prediction apparatus, including a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps according to the above-described method.
In a fourth aspect, the invention provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the above-described method.
Compared with the prior art, the invention has the following beneficial effects:
according to the power load prediction method, the device and the storage medium, the power load prediction model is constructed by utilizing the convolutional neural network and combining the attention mechanism to perform offline regression learning, so that the weight of important information can be improved, the weight of useless information can be reduced, the offline learning efficiency can be improved, and the complexity of the model can be reduced; the method can convert one-dimensional data into two-dimensional images, thereby being applicable to most deep learning networks and expanding the application range; the constructed image is preprocessed by utilizing an image processing technology, and a data matrix is expanded through cubic interpolation, so that more characteristic expressions are obtained, the offline learning efficiency is improved, and the model prediction accuracy is improved.
Drawings
Fig. 1 is a flowchart of a power load prediction method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating a transition of a cyclic shift method according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of transforming a gray image into an RGB image according to an embodiment of the present invention;
FIG. 4 is a schematic diagram of a process for constructing a power load prediction model according to an embodiment of the present invention;
FIG. 5 is a schematic diagram of a convolutional neural network structure incorporating an attention module according to an embodiment of the present invention;
FIG. 6 is a schematic view of an attention module according to an embodiment of the present invention;
fig. 7 is a graph depicting predicted performance of an electrical load according to an embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
The first embodiment is as follows:
as shown in fig. 1, an embodiment of the present invention provides a power load prediction method, including the following steps:
101. and acquiring meteorological data and corresponding time information on line.
102. Preprocessing the online acquired meteorological data and corresponding time information to generate online comprehensive data; the process is as follows:
meteorological data: in the specific implementation process, 4 different types of meteorological data of temperature, pressure, wind speed and humidity are collected through a meteorological sensor; because meteorological data originates from different sensors and the measurements are not of the same size. In order to remove the influence of dimension and increase the comparability of data, each meteorological data is normalized:
in the formula, x i,n 、x′ i,n Respectively the value, x, of the ith meteorological data before and after normalization i,max 、x i,min The maximum value and the minimum value of the ith meteorological data are obtained; i =1, 2, 3, 4; each meteorological datum is shifted by a range of 0-1 through normalization.
Time information: twelve month portions in one year are coded according to the month portions, and the codes are 1-12; twenty-four hours of the day are coded as 1, 2, 3, 4 according to 0-6, 6-12, 12-18, 18-24, respectively.
And (3) online comprehensive data: merging the normalized meteorological data and the encoded time information:
x n ＝[x 1,n ,…,x m,n ,t n ,d n ]
in the formula, x n For the nth online integration data, x m,n Is the m-th meteorological data, t n And d n The codes are respectively corresponding to the number of months and the number of hours.
103. Converting the online comprehensive data into an online data image based on an image processing technology; the method specifically comprises the following steps:
s1, converting online comprehensive data into an online comprehensive data matrix through a cyclic shift method;
as shown in fig. 2, the transition process is as follows:
(1) And taking the online comprehensive data as a first row of an online comprehensive data matrix.
(2) The i, i >1 th row of the online synthetic data matrix has elements in the i-1 th row shifted left by one bit, and the first element is placed in the last bit of the row.
(3) Cyclically shifted q-1 times according to a given number of rows q, to form an on-line synthetic data matrix having a size of (m + 1) × q.
S2, carrying out interpolation processing on the online comprehensive data matrix through a cubic spline interpolation method to complete the dimension expansion of the online comprehensive data matrix; the principle of the cubic spline interpolation method is as follows:
given n points, a = y 0 <y 1 <…<y = b, and respectively knows the function values f (y) corresponding to them i ) The next cubic polynomial is determined to predict the values of the other points.
S i (y)＝a i +b i (y-y i )+c i (y-y i ) 2 +d i (y-y i ) 3
Each cubic polynomial satisfies the following condition:
1.S i (y) second order conductibility
2.S i (y) in the interval (y) 0 ，y n ) Continuous
3.S′ i (y) in the interval (y) 0 ，y n ) Continuous
4.S″ i (y) in the interval (y) 0 ，y n ) Continuous
Given a fixed boundary condition S' i (y 0 )＝z′ 0 ，S′ n-1 (y n )＝z′ n
Given a natural boundary condition S ″ i (y 0 )＝z″ 0 ＝0，S″ n-1 (xy n )＝z″ n ＝0
Can be obtained by simultaneous equations to obtain a i b i c i d i The value of (c). Derived S i (y) new interpolated data can be obtained by substituting points that increment y.
And S3, mapping the element values in the online comprehensive data matrix after the dimensionality expansion into pixel values in an RGB image through a linear mapping method to form an online data image.
In this embodiment, each position element value in the online integrated data matrix corresponds to a gray pixel value at a corresponding position of the gray map. The size of the gray image is consistent with the size of the two-dimensional matrix. Since the elements in the online aggregated data matrix are all between 0-1, if the selected gray scale range is 256 gray scale orders of magnitude. We multiply the matrix elements by 256 and then correspond to one gray value each, thus constructing the complete online data image.
Finally, the gray pixel values in the gray image are corresponding to each color by utilizing a gray-color conversion method, as shown in fig. 3, the gray values in the gray image are respectively subjected to three different conversions of red conversion, green conversion and blue conversion to form three pixel values of RGB. And then synthesizing the color image.
The red transform, the green transform, and the blue transform utilized by this patent for each gray level pixel are defined as follows:
104. and inputting the online data image into a pre-constructed power load prediction model to obtain a power load prediction value.
As shown in fig. 4, the process of constructing the power load prediction model includes:
201. acquiring meteorological data, corresponding time information and an actual value of a power load offline;
202. preprocessing the meteorological data acquired offline and the corresponding time information to generate offline comprehensive data; the way of preprocessing here is the same principle as step 102.
203. Converting the offline comprehensive data into an offline data image based on an image processing technology; the transformation here is in the same manner as in step 103.
204. And constructing a convolutional neural network combined with an attention module, and training the offline data image and the corresponding power load actual value respectively as the input and the output of the convolutional neural network to generate a power load prediction model.
As shown in fig. 5, the convolutional neural network includes a first convolutional layer, a pooling layer, an attention module, a second convolutional layer, a fully-connected layer, and an output layer, which are connected in sequence; the first convolution layer and the second convolution layer respectively adopt 32 convolution kernels and 64 convolution kernels with the size of 3 x 3, and the ReLU function is selected as the activation function; selecting a maximum pooling method for the pooling layer, setting the size of the pooling layer to be 2 multiplied by 2, uniformly selecting an SAME mode for the filling mode, and taking the step length to be 1; the full-connection layer is constructed by 64 neurons, and the ReLU function is selected by the activation function. And in the output layer, activating a function to select a linear function, and completing load prediction based on the linear function.
As shown in fig. 6, in the present embodiment, a lightweight Attention Module CBAM (conditional Block Attention Module) is optionally added to the Convolutional neural network CNN, and this Module combines the spatial and channel Attention mechanisms to extract the Attention of the channel and the spatial Attention respectively, so as to achieve a better effect. The attention module combines the channel attention module and the spatial attention module in a serial fashion.
As shown in fig. 7, it can be found that the actual value and the predicted value can be substantially matched by statistically plotting the actual value and the predicted value of the power load. Through statistical analysis of the error, the average estimated error for this technique is 225. Therefore, the method and the device can accurately predict the power load data and meet actual requirements very well.
Example two:
the embodiment of the invention provides a power load prediction device, which comprises:
the online data module is used for acquiring meteorological data and corresponding time information online;
the online preprocessing module is used for preprocessing the online acquired meteorological data and the corresponding time information to generate online comprehensive data;
the online conversion module is used for converting the online comprehensive data into an online data image based on an image processing technology;
the online prediction module is used for inputting online data images into a pre-constructed power load prediction model to obtain a power load prediction value;
the construction process of the power load prediction model comprises the following steps:
the off-line data module is used for off-line acquiring meteorological data, corresponding time information and an actual value of the power load;
the off-line preprocessing module is used for preprocessing the off-line acquired meteorological data and the corresponding time information to generate off-line comprehensive data;
the offline conversion module is used for converting the offline comprehensive data into an offline data image based on an image processing technology;
and the model building module is used for building a convolutional neural network combined with the attention module, and training the offline data image and the corresponding power load actual value as the input and the output of the convolutional neural network respectively to generate a power load prediction model.
Specifically, the step of preprocessing the online acquired meteorological data and the corresponding time information to generate online comprehensive data includes:
meteorological data: normalization was performed for each meteorological datum:
in the formula, x i,n 、x′ i,n Respectively the value, x, of the ith meteorological data before and after normalization i,max 、x i,min The maximum value and the minimum value of the ith meteorological data are obtained;
time information: twelve months in a year are coded according to the number of months, and twenty-four hours in a day are respectively coded as 1, 2, 3 and 4 according to 0-6, 6-12, 12-18 and 18-24;
and (3) online comprehensive data: merging the normalized meteorological data and the encoded time information:
x n ＝[x 1,n ,…,x m,n ,t n ,d n ]
in the formula, x n For the nth online integration data, x m,n Is the m-th meteorological data, t n And d n The codes are respectively corresponding to the number of months and the number of hours.
Specifically, converting the online integrated data into an online data image based on an image processing technique includes:
converting the online comprehensive data into an online comprehensive data matrix by a cyclic shift method;
carrying out interpolation processing on the online comprehensive data matrix by a cubic spline interpolation method to complete the dimension expansion of the online comprehensive data matrix;
and mapping the element values in the online comprehensive data matrix after the dimensionality expansion into pixel values in an RGB image by a linear mapping method to form an online data image.
Specifically, the convolutional neural network comprises a first convolutional layer, a pooling layer, an attention module, a second convolutional layer, a full-link layer and an output layer which are sequentially connected; the attention module includes a channel attention module and a spatial attention module combined in a serial manner.
Example three:
based on the first embodiment, the embodiment of the invention provides a power load prediction device, which comprises a processor and a storage medium;
a storage medium to store instructions;
the processor is configured to operate in accordance with instructions to perform steps in accordance with the above-described method.
Example four:
according to a first embodiment, an embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of the method.
The method utilizes two important factors, namely time information and meteorological data, which influence the power load prediction to carry out regression learning through an offline learning network of a CNN + attention mechanism, so as to realize the power load prediction.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application 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, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams 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.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.
Claims (10)
1. A method for predicting a power load, comprising:
acquiring meteorological data and corresponding time information on line;
preprocessing the online acquired meteorological data and corresponding time information to generate online comprehensive data;
converting the online comprehensive data into an online data image based on an image processing technology;
inputting an online data image into a pre-constructed power load prediction model to obtain a power load prediction value;
wherein, the construction process of the power load prediction model comprises the following steps:
acquiring meteorological data, corresponding time information and an actual value of a power load off line;
preprocessing the meteorological data acquired offline and the corresponding time information to generate offline comprehensive data;
converting the offline comprehensive data into an offline data image based on an image processing technology;
and constructing a convolutional neural network combined with an attention module, and training the offline data image and the corresponding power load actual value respectively as the input and the output of the convolutional neural network to generate a power load prediction model.
2. The method of claim 1, wherein the preprocessing the online acquired meteorological data and the corresponding time information to generate online comprehensive data comprises:
meteorological data: normalization was performed for each meteorological datum:
in the formula, x i,n 、x′ i,n Respectively the value, x, of the ith meteorological data before and after normalization i,max 、x i,min The maximum value and the minimum value of the ith meteorological data are obtained;
time information: twelve months in a year are coded according to the number of months, and twenty-four hours in a day are respectively coded as 1, 2, 3 and 4 according to 0-6, 6-12, 12-18 and 18-24;
and (3) online comprehensive data: merging the normalized meteorological data and the encoded time information:
x n ＝[x 1,n ,…,x m,n ,t n ,d n ]
in the formula, x n For the nth online integration data, x m,n Is the m-th meteorological data, t n And d n The number of months and hours respectively.
3. The method of claim 1, wherein the converting the online synthetic data into an online data image based on an image processing technique comprises:
converting the online comprehensive data into an online comprehensive data matrix by a cyclic shift method;
carrying out interpolation processing on the online comprehensive data matrix by a cubic spline interpolation method to complete the dimension expansion of the online comprehensive data matrix;
and mapping the element values in the dimension-expanded online comprehensive data matrix into pixel values in an RGB image by a linear mapping method to form an online data image.
4. The power load prediction method according to claim 1, wherein the convolutional neural network comprises a first convolutional layer, a pooling layer, an attention module, a second convolutional layer, a fully-connected layer and an output layer which are connected in sequence; the attention module includes a channel attention module and a spatial attention module combined in a serial manner.
5. An electrical load prediction apparatus, the apparatus comprising:
the online data module is used for acquiring meteorological data and corresponding time information online;
the online preprocessing module is used for preprocessing the online acquired meteorological data and the corresponding time information to generate online comprehensive data;
the online conversion module is used for converting the online comprehensive data into an online data image based on an image processing technology;
the online prediction module is used for inputting online data images into a pre-constructed power load prediction model to obtain a power load prediction value;
wherein, the construction process of the power load prediction model comprises the following steps:
the off-line data module is used for off-line acquiring meteorological data, corresponding time information and an actual value of the power load;
the off-line preprocessing module is used for preprocessing the off-line acquired meteorological data and the corresponding time information to generate off-line comprehensive data;
the off-line conversion module is used for converting the off-line comprehensive data into an off-line data image based on an image processing technology;
and the model building module is used for building a convolutional neural network combined with the attention module, and training the offline data image and the corresponding power load actual value as the input and the output of the convolutional neural network respectively to generate a power load prediction model.
6. The power load prediction device of claim 1, wherein the pre-processing the online acquired meteorological data and corresponding time information to generate online integrated data comprises:
meteorological data: normalization was performed for each meteorological datum:
in the formula, x i,n 、x′ i,n Respectively the value, x, of the ith meteorological data before and after normalization i,max 、x i,min The maximum value and the minimum value of the ith meteorological data are obtained;
time information: twelve month portions in one year are coded according to the number of months, and twenty-four hours in one day are respectively coded into 1, 2, 3 and 4 according to 0-6, 6-12, 12-18 and 18-24;
and (3) online comprehensive data: merging the normalized meteorological data and the encoded time information:
x n ＝[x 1,n ,…,x m,n ,t n ,d n ]
in the formula, x n For the nth online integration data, x m,n Is the m-th meteorological data, t n And d n The codes are respectively corresponding to the number of months and the number of hours.
7. The power load prediction device of claim 1, wherein the converting the online synthetic data into an online data image based on an image processing technique comprises:
converting the online comprehensive data into an online comprehensive data matrix by a cyclic shift method;
carrying out interpolation processing on the online comprehensive data matrix by a cubic spline interpolation method to complete the dimension expansion of the online comprehensive data matrix;
and mapping the element values in the online comprehensive data matrix after the dimensionality expansion into pixel values in an RGB image by a linear mapping method to form an online data image.
8. The power load prediction device of claim 1, wherein the convolutional neural network comprises a first convolutional layer, a pooling layer, an attention module, a second convolutional layer, a fully-connected layer and an output layer which are connected in sequence; the attention module includes a channel attention module and a spatial attention module combined in a serial manner.
9. An electrical load prediction apparatus comprising a processor and a storage medium;
the storage medium is used for storing instructions;
the processor is configured to operate in accordance with the instructions to perform the steps of the method according to any one of claims 1 to 4.
10. Computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210710222.7A CN115204467A (en) | 2022-06-22 | 2022-06-22 | Power load prediction method, device and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210710222.7A CN115204467A (en) | 2022-06-22 | 2022-06-22 | Power load prediction method, device and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115204467A true CN115204467A (en) | 2022-10-18 |
Family
ID=83577004
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210710222.7A Pending CN115204467A (en) | 2022-06-22 | 2022-06-22 | Power load prediction method, device and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115204467A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115937355A (en) * | 2023-02-14 | 2023-04-07 | 国网湖北省电力有限公司电力科学研究院 | Power Internet of things load prediction fingerprint image construction method, system and medium |
-
2022
- 2022-06-22 CN CN202210710222.7A patent/CN115204467A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115937355A (en) * | 2023-02-14 | 2023-04-07 | 国网湖北省电力有限公司电力科学研究院 | Power Internet of things load prediction fingerprint image construction method, system and medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113177633B (en) | Depth decoupling time sequence prediction method | |
| CN111680841B (en) | Short-term load prediction method, system and terminal equipment based on principal component analysis | |
| CN116845889B (en) | Hierarchical hypergraph neural network-based power load prediction method | |
| CN117096867A (en) | Short-term power load prediction method, device, system and storage medium | |
| CN115600640A (en) | Power load prediction method based on decomposition network | |
| CN117216526B (en) | Photovoltaic output prediction method and system based on artificial intelligence | |
| CN116865261B (en) | Power load prediction method and system based on twin network | |
| CN112508286A (en) | Short-term load prediction method based on Kmeans-BilSTM-DMD model | |
| CN115658671B (en) | Container port loading and unloading equipment fault data complement method based on deep learning | |
| CN115204467A (en) | Power load prediction method, device and storage medium | |
| JP7570747B1 (en) | A forecasting platform for forecasting time series data of energy amounts | |
| CN115034430A (en) | Carbon emission prediction method, device, terminal and storage medium | |
| CN116777668A (en) | Multi-energy load prediction method and system based on improved coding and decoding model | |
| CN114091776A (en) | K-means-based multi-branch AGCNN short-term power load prediction method | |
| CN116014722A (en) | Sub-solar photovoltaic power generation prediction method and system based on seasonal decomposition and convolution network | |
| CN118036822A (en) | Electric hydrogen load prediction method, device and storage medium of electric hydrogen integrated energy station | |
| JP2024156234A (en) | A forecasting platform for forecasting time series data of energy amounts | |
| CN117407681B (en) | Time sequence data prediction model establishment method based on vector clustering | |
| CN117874474A (en) | Deep learning-based charge station idle state equipment quantity prediction method | |
| CN117436920A (en) | Hybrid neural network day-ahead electricity price prediction method embedded with multi-head attention mechanism | |
| CN117498296A (en) | Multi-wind power plant power prediction method based on attention space-time synchronization diagram convolution network | |
| CN117113054A (en) | Multi-element time sequence prediction method based on graph neural network and transducer | |
| CN116402194A (en) | Multi-time scale load prediction method based on hybrid neural network | |
| CN113095596B (en) | Photovoltaic power prediction method based on multi-stage Gate-SA-TCN | |
| CN109472366A (en) | A kind of coding-decoding method and device of machine learning model |
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 |