CN116822739A - Medium-long-term sales electricity quantity prediction method, device, terminal and storage medium - Google Patents
Medium-long-term sales electricity quantity prediction method, device, terminal and storage medium Download PDFInfo
- Publication number
- CN116822739A CN116822739A CN202310797117.6A CN202310797117A CN116822739A CN 116822739 A CN116822739 A CN 116822739A CN 202310797117 A CN202310797117 A CN 202310797117A CN 116822739 A CN116822739 A CN 116822739A
- Authority
- CN
- China
- Prior art keywords
- prediction
- sales
- medium
- long
- term
- 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
- 230000005611 electricity Effects 0.000 title claims abstract description 84
- 238000000034 method Methods 0.000 title claims abstract description 57
- 238000003860 storage Methods 0.000 title claims abstract description 16
- 238000005457 optimization Methods 0.000 claims abstract description 85
- 238000012545 processing Methods 0.000 claims abstract description 23
- 230000009467 reduction Effects 0.000 claims abstract description 10
- 238000004590 computer program Methods 0.000 claims description 23
- 239000011159 matrix material Substances 0.000 claims description 20
- 230000006870 function Effects 0.000 claims description 18
- 230000008569 process Effects 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 15
- 230000007774 longterm Effects 0.000 claims description 4
- 230000008859 change Effects 0.000 abstract description 7
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 26
- 229910000831 Steel Inorganic materials 0.000 description 14
- 239000010959 steel Substances 0.000 description 14
- 229910052742 iron Inorganic materials 0.000 description 13
- 238000010586 diagram Methods 0.000 description 4
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 3
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 238000007726 management method Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000005612 types of electricity Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a medium-and-long-term electricity sales quantity prediction method, a medium-and-long-term electricity sales quantity prediction device, a terminal and a storage medium. The method comprises the following steps: acquiring electricity sales volume sample data under different types in a power grid, performing dimension reduction processing on the electricity sales volume sample data to obtain feature data corresponding to different user groups under different time dimensions, inputting the feature data into a preset prediction model library to obtain electricity sales volume prediction reference values, inputting the feature data and the electricity sales volume prediction reference values into an optimization model to obtain optimization parameters corresponding to the feature data of different user groups under different time dimensions, adjusting the preset prediction model library according to the optimization parameters to obtain an optimized prediction model library, inputting the feature data into the optimized prediction model library, and performing medium-long-term prediction on the electricity sales volume under different types to obtain the medium-long-term electricity sales volume prediction values of the power grid. The application takes the electricity utilization characteristics of different user groups under different time dimensions into consideration, and accurately predicts the sales power quantities with different change rules.
Description
Technical Field
The application relates to the technical field of power systems, in particular to a medium-long-term electricity sales quantity prediction method, a medium-long-term electricity sales quantity prediction device, a terminal and a storage medium.
Background
Along with the continuous increase of the electricity consumption scale of China, the influence of the difference and uncertainty fluctuation of the electricity consumption of different industries on the power grid dispatching and the safe operation is continuously deepened, so that the prediction of the medium-and-long-term electricity sales of different industries is very necessary.
At present, most of traditional prediction methods utilize linear relations between influence factors and prediction targets to predict medium and long-term sales power, and nonlinear characteristics between sales power and various influence factors cannot be considered. The artificial intelligent prediction method can process nonlinear related factors, has good robustness, but is difficult to process large-scale data and has poor data sensitivity.
Therefore, how to accurately predict the sales power with different change rules is a problem to be solved in the present day.
Disclosure of Invention
The embodiment of the application provides a medium-and-long-term electricity sales quantity prediction method, a device, a terminal and a storage medium, which are used for solving the problem that the prior art cannot accurately predict electricity sales quantities with different change rules.
In a first aspect, an embodiment of the present application provides a method for predicting medium-to-long-term electricity sales amount, including:
acquiring electricity sales quantity sample data under different types in a power grid; wherein the type includes a plurality of user groups and a plurality of time dimensions;
performing dimension reduction processing on the electricity sales volume sample data to obtain characteristic data corresponding to different user groups in different time dimensions;
inputting the characteristic data into a preset prediction model library to obtain a sales quantity prediction reference value;
inputting the characteristic data and the sales quantity prediction reference value into an optimization model to obtain optimization parameters corresponding to the characteristic data of different user groups under different time dimensions;
adjusting a preset prediction model library according to the optimization parameters to obtain an optimized prediction model library;
and inputting the characteristic data into an optimized prediction model library, and performing medium-long-term prediction on the sales power under different types to obtain a medium-long-term sales power predicted value of the power grid.
In one possible implementation manner, the optimization model includes a plurality of full-connection layers, and the feature data and the sales quantity prediction reference value are input into the optimization model to obtain optimization parameters corresponding to the feature data of different user groups under different time dimensions, including:
calculating a first average value corresponding to the characteristic data and a second average value corresponding to the sales quantity prediction reference value;
obtaining a distance reference value based on the difference between the first average value and the second average value;
initializing the difference distance between two adjacent full-connection layers in the optimization model according to the distance reference value to obtain an initialized optimization model;
and obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the initialized optimization model.
In one possible implementation manner, after inputting the feature data into a preset prediction model library to obtain the sales quantity prediction reference value, the method further includes:
acquiring a first loss value; the first loss value is an error between the characteristic data and the sales quantity prediction reference value;
according to the initialized optimization model, obtaining optimization parameters corresponding to the feature data of different user groups under different time dimensions, wherein the optimization parameters comprise:
inputting the characteristic data into the initialized optimization model to predict the sales amount;
calculating the difference distance between two adjacent full-connection layers in the initialized optimization model in the prediction process;
constructing an objective function according to the first loss value and the difference distance;
obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the minimum value of the objective function; wherein the optimization parameters include depth and width of the fully connected layer.
In one possible implementation manner, performing dimension reduction processing on the electricity sales volume sample data to obtain feature data corresponding to different user groups in different time dimensions, where the feature data includes:
calculating a corresponding global divergence matrix according to the electricity sales volume sample data;
obtaining a projection matrix corresponding to the sales quantity sample data according to the global divergence matrix;
and obtaining characteristic data corresponding to different user groups under different time dimensions according to the projection matrix and the sales volume sample data.
In one possible implementation, the preset library of prediction models includes a plurality of different prediction models.
In one possible implementation manner, each prediction model is provided with a tag, the type of the tag corresponds to the type of the electricity sales volume sample data, the feature data is input into a preset prediction model library, and the electricity sales volume prediction reference value is obtained, including:
searching a prediction model matched with the feature type in a preset prediction model library based on the feature type of the feature data to obtain a reference sequence; wherein the feature type corresponds to the type of the tag;
calculating the weight of each prediction model in the reference sequence, and determining the prediction model with the largest weight in the reference sequence as a reference model;
and inputting the characteristic data into a reference model to obtain a sales quantity prediction reference value.
In one possible implementation manner, the adjusting the preset prediction model library according to the optimization parameters to obtain an optimized prediction model library includes:
adjusting a reference model in a preset prediction model library based on the optimization parameters to obtain an optimized reference model;
inputting the characteristic data into an optimized prediction model library, and performing medium-long-term prediction on the sales power under different types to obtain a medium-long-term sales power predicted value, wherein the method comprises the following steps:
and inputting the characteristic data into the optimized reference model to obtain the medium-and-long-term sales quantity predicted value corresponding to the characteristic type of the characteristic data.
In a second aspect, an embodiment of the present application provides a medium-and-long-term electricity sales amount prediction apparatus, including:
the acquisition module is used for acquiring the electricity sales quantity sample data under different types in the power grid; wherein the type includes a plurality of user groups and a plurality of time dimensions;
the data processing module is used for performing dimension reduction processing on the electricity sales volume sample data to obtain characteristic data corresponding to different user groups under different time dimensions;
the reference value calculation module is used for inputting the characteristic data into a preset prediction model library to obtain a sales quantity prediction reference value;
the parameter calculation module is used for inputting the characteristic data and the sales quantity prediction reference value into the optimization model to obtain optimization parameters corresponding to the characteristic data of different user groups under different time dimensions;
the optimization module is used for adjusting a preset prediction model library according to the optimization parameters to obtain an optimized prediction model library;
and the prediction module is used for inputting the characteristic data into the optimized prediction model library, and performing medium-long-term prediction on the sales power under different types to obtain a medium-long-term sales power predicted value of the power grid.
In a third aspect, an embodiment of the present application provides a terminal, a memory and a processor, where the memory is configured to store a computer program, and the processor is configured to call and run the computer program stored in the memory, and when the processor executes the computer program, implement the steps of the medium-long term electricity sales prediction method according to the first aspect or any one of possible implementation manners of the first aspect.
In a fourth aspect, an embodiment of the present application provides a computer readable storage medium storing a computer program, which when executed by a processor implements the steps of the medium-to-long term sales volume prediction method according to the first aspect or any one of the possible implementations of the first aspect.
The embodiment of the application provides a medium-and-long-term electricity sales quantity prediction method, a device, a terminal and a storage medium. The characteristic data are input into a preset prediction model library to obtain a sales quantity prediction reference value, and further, through an optimization model, optimization parameters corresponding to the characteristic data of different user groups in different time dimensions are obtained, the preset prediction model library is adjusted according to the optimization parameters to obtain an optimized prediction model library, so that the sales quantity of different user groups in different time dimensions is predicted for a medium period and a long period to obtain the medium-period sales quantity prediction value of the power grid. According to the embodiment of the application, the electricity utilization characteristics of different user groups under different time dimensions are taken into consideration, and the electricity sales quantity with different change rules can be accurately predicted.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments or the description of the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of an implementation of a method for predicting medium-to-long-term sales power according to an embodiment of the present application;
FIG. 2 is a flowchart of an implementation of an optimization model initialization method provided by an embodiment of the present application;
FIG. 3 is a flowchart of an implementation of an optimization parameter acquisition method provided by an embodiment of the present application;
fig. 4 is a schematic structural diagram of a medium-to-long-term electricity sales prediction device according to an embodiment of the present application;
fig. 5 is a schematic diagram of a terminal according to an embodiment of the present application;
FIG. 6 (a) is a comparison plot of NRMSE of a predicted outcome provided by an embodiment of the present application;
FIG. 6 (b) is a comparison map of MAPE of a predicted result provided by an embodiment of the present application.
Detailed Description
In the following description, for purposes of explanation and not limitation, specific details are set forth such as the particular system architecture, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. It will be apparent, however, to one skilled in the art that the present application may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.
Along with the continuous increase of the electricity consumption scale of China, the influence of the difference and uncertainty fluctuation of the electricity consumption of different industries on the power grid dispatching and the safe operation is continuously deepened, and the accurate prediction of the electricity sales of different industries is one of the important means for solving the problems. The medium-and-long-term electricity sales prediction is a basis for power grid operation scheduling and power demand side management, and when the power consumption environment is increasingly complex, the medium-and-long-term electricity sales prediction work is carried out, so that the reasonable planning of a power grid is facilitated, and the intelligent and lean levels of power grid management are improved.
At present, most of traditional prediction methods utilize linear relations between influence factors and prediction targets to predict medium-long-term electricity sales, and the prediction performance of a model is improved by combining multiple regression ideas, but nonlinear characteristics between the electricity sales and various influence factors cannot be considered. The artificial intelligent prediction method can process nonlinear related factors, has good robustness, but is difficult to process large-scale data, cannot better reflect data characteristics, and has poor data sensitivity.
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the following description will be made by way of specific embodiments with reference to the accompanying drawings.
In order to solve the problem that the prior art cannot accurately predict the sales power with different change rules, the embodiment of the application provides a middle-long-term sales power prediction method, a device, a terminal and a storage medium.
Fig. 1 is a flowchart of an implementation of a method for predicting medium-to-long-term electricity sales according to an embodiment of the present application, which is described in detail below:
step 101, acquiring electricity sales quantity sample data under different types in a power grid; wherein the type includes a plurality of user groups and a plurality of time dimensions.
In some embodiments, the electricity sales volume sample data of the electric network in two years can be used as a data base, and the user group of the electric network mainly comprises a resident population and industrial users, wherein the industrial users further comprise industries such as steel industry, chemical industry, raw material production and the like.
And 102, performing dimension reduction processing on the electricity sales volume sample data to obtain feature data corresponding to different user groups under different time dimensions.
In some embodiments, the electricity sales volume sample data may be clustered by maximizing an average value between classes and minimizing an intra-class variance, and the electricity sales volume sample data is projected onto a low latitude, so that different types of electricity sales volume sample data are classified according to different time dimensions and different user groups, and feature data including electricity consumption characteristics of users are obtained.
In some embodiments, the specific processing of step 102 may be as follows:
firstly, according to the sales quantity sample data, a corresponding global divergence matrix is calculated.
Specifically, the global divergence matrix may be calculated according to the following formula, namely:
wherein S is t Representing a global divergence matrix, x i Represents the ith sample data, μ represents the mean vector of all samples.
And secondly, obtaining a projection matrix corresponding to the sales quantity sample data according to the global divergence matrix.
Specifically, the eigenvalues and eigenvectors of the global divergence matrix can be calculated, and the first d eigenvalues and the corresponding d eigenvectors (w 1 ,w 2 ,...,w d ) A projection matrix W is obtained.
And finally, obtaining the characteristic data corresponding to different user groups under different time dimensions according to the projection matrix and the sales volume sample data.
Specifically, feature data corresponding to different user groups in different time dimensions can be obtained according to the following formula, namely:
z i ＝W T x i ；
wherein z is i Representing the resulting characteristic data.
And step 103, inputting the characteristic data into a preset prediction model library to obtain a sales quantity prediction reference value.
In some embodiments, the preset prediction model library comprises a plurality of different prediction models, and each prediction model is provided with a label, and the types of the labels correspond to the types of the sales quantity sample data.
Specifically, the preset prediction model library may include a time sequence prediction model, a gray level algorithm prediction model, an LSTM prediction model, an RNN prediction model, or a CNN prediction model. The time sequence prediction model has better time sequence feature extraction capability, and can better extract the fluctuation rule of the electricity sales quantity sample data on the time sequence; the gray algorithm prediction model has low requirements on data integrity and reliability, and can fully mine the essential characteristics of data by utilizing a differential equation, so that the situation of future development trend of things is predicted; the LSTM prediction model can selectively store information, prolong the time sequence information learning range of the neural network, and is suitable for predicting medium-and-long-term electricity sales under different user groups with more data; the RNN predictive model has significant advantages when learning the nonlinear characteristics of data.
Taking the prediction of sales power in different time dimensions as an example, since the time series model is more accurate for long-term sales power prediction, a label of "two months" or "one month" can be set on the time series prediction model.
In some embodiments, the specific processing of step 103 may be as follows:
firstly, searching a prediction model matched with a feature type in a preset prediction model library based on the feature type of feature data to obtain a reference sequence; wherein the feature type corresponds to the type of tag.
Specifically, taking the example of predicting the monthly electricity sales of the iron and steel industry, a prediction model provided with an iron and steel tag and a one month tag can be searched to obtain a corresponding reference sequence.
And secondly, calculating the weight of each prediction model in the reference sequence, and determining the prediction model with the largest weight in the reference sequence as the reference model.
Specifically, a weight calculation method can be set in combination with actual production requirements, the weight of each prediction model in the reference sequence is calculated, and the prediction model with the largest weight is set as the reference model so as to predict the sales amount of the type.
And finally, inputting the characteristic data into a reference model to obtain a sales quantity prediction reference value.
Specifically, taking the example of predicting the monthly electricity sales of the iron and steel industry, the corresponding characteristic data is input into the reference model to obtain the predicted reference value of the monthly electricity sales of the iron and steel industry.
In some embodiments, after the feature data is input into a preset prediction model library to obtain the sales power prediction reference value, a first loss value is further obtained, where the first loss value is an error between the feature data and the sales power prediction reference value.
And 104, inputting the characteristic data and the sales quantity prediction reference value into an optimization model to obtain optimization parameters corresponding to the characteristic data of different user groups in different time dimensions.
In some embodiments, the optimization model includes multiple fully connected layers, as shown in FIG. 2, the specific process of step 104 may be as follows:
step 201, calculating a first average value corresponding to the feature data and a second average value corresponding to the sales quantity prediction reference value.
In some embodiments, taking the example of predicting the monthly sales power of the iron and steel industry, a first average corresponding to the characteristic data of the iron and steel industry per month and a second average corresponding to the predicted reference value of the monthly sales power of the iron and steel industry may be calculated.
Step 202, obtaining a distance reference value based on the difference between the first average value and the second average value.
In some embodiments, taking the example of predicting the monthly electricity sales of the iron and steel industry, a difference between a first average value corresponding to the characteristic data of the iron and steel industry and a second average value corresponding to the predicted reference value of the monthly electricity sales of the iron and steel industry may be calculated to obtain the distance reference value.
And 203, initializing the difference distance between two adjacent full-connection layers in the optimization model according to the distance reference value to obtain an initialized optimization model.
In some embodiments, taking prediction of monthly sales power of the iron and steel industry as an example, parameters in the selected reference model are initialized according to the obtained distance reference value, so as to obtain an initialized reference model.
And 204, obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the initialized optimization model.
In some embodiments, taking the example of predicting monthly sales power of the iron and steel industry, the initialized reference model is used to predict sales power, so as to obtain optimized parameters suitable for the reference model.
In some embodiments, as shown in FIG. 3, the specific process of step 204 may be as follows:
and step 301, inputting the characteristic data into the initialized optimization model to predict the sales amount.
In some embodiments, taking the example of predicting the monthly sales capacity of the iron and steel industry, corresponding characteristic data is input into an initialized reference model to predict the sales capacity.
Step 302, calculating a difference distance between two adjacent fully connected layers in the initialized optimization model in the prediction process.
In some embodiments, the sum of the difference distances at a given layer may be calculated according to the following formula:
wherein lambda represents an importance parameter,representing the sum of the difference distances (MDD, margin Disparity Discrepancy) on the first layer. It should be noted that the difference distance may determine the depth and width of the feature layer in the neural network during training.
Step 303, constructing an objective function according to the first loss value and the difference distance.
In some embodiments, the objective function may be constructed according to the following formula, namely:
L＝L c (X s ,y)+L 2 ；
wherein L is c (X s Y) represents a first loss value, i.e. the error between the characteristic data and the sales quantity prediction reference value, X s And the characteristic data is represented, and y represents a sales quantity prediction reference value.
Step 304, obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the minimum value of the objective function; wherein the optimization parameters include depth and width of the fully connected layer.
In some embodiments, the objective function is continuously trained in the process of sales power prediction, and after the training is finished, the minimum value of the objective function is taken, and the optimization parameter at the moment is obtained according to the minimum value of the objective function.
And 105, adjusting a preset prediction model library according to the optimization parameters to obtain an optimized prediction model library.
In some embodiments, the reference model in the preset prediction model library may be adjusted based on the optimization parameters, so as to obtain an optimized reference model.
And 106, inputting the characteristic data into an optimized prediction model library, and performing medium-long-term prediction on the sales power under different types to obtain a medium-long-term sales power predicted value of the power grid.
In some embodiments, the feature data may be input into an optimized reference model to obtain a medium-to-long-term sales volume prediction value corresponding to the feature type of the feature data.
Taking sales amount sample data with different time dimensions as an example, data with different time lengths are respectively input into the model library for prediction, and the prediction accuracy of each model is evaluated by using two indexes of NRMSE and MAPE, see FIG. 6 (a) and FIG. 6 (b), wherein NRMSE and MAPE can be calculated according to the following formulas:
wherein, the liquid crystal display device comprises a liquid crystal display device,indicating the predicted value of sales power->For the actual value of the sales power at time t, < >>And T is the time sequence length, and is the maximum value of the sales electricity quantity in the period.
As shown in fig. 6 (a) and fig. 6 (b), compared with a single model prediction result, the optimized prediction model library in the embodiment of the application can better track the fluctuation characteristics of the sales power curve, the prediction result is more stable, and the middle-long-term prediction result is more accurate.
The embodiment of the application provides a medium-and-long-term electricity sales prediction method, which is characterized in that characteristic data corresponding to different user groups in different time dimensions are obtained by acquiring electricity sales volume sample data in a power grid and performing dimension reduction processing on the electricity sales volume sample data, so that different electricity utilization characteristics are classified. And inputting the characteristic data into a preset prediction model library to obtain a sales quantity prediction reference value, further obtaining optimization parameters corresponding to the characteristic data of different user groups in different time dimensions through an optimization model, and adjusting the preset prediction model library according to the optimization parameters so as to predict the sales quantity of different user groups in different time dimensions for medium and long periods and obtain a medium and long-term sales quantity prediction value of the power grid. The embodiment of the application can take the electricity utilization characteristics of different user groups under different time dimensions into consideration, and accurately predict the sales power quantities with different change rules.
It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiment of the present application.
The following are device embodiments of the application, for details not described in detail therein, reference may be made to the corresponding method embodiments described above.
Fig. 4 is a schematic structural diagram of a medium-to-long-term electricity sales prediction device according to an embodiment of the present application, and for convenience of explanation, only the relevant parts of the embodiment of the present application are shown, which is described in detail below:
as shown in fig. 4, the medium-to-long-term sales amount prediction apparatus 4 includes: an acquisition module 401, a data processing module 402, a reference value calculation module 403, a parameter calculation module 404, an optimization module 405 and a prediction module 406.
The acquisition module 401 is configured to acquire sales amount sample data under different types in the power grid; wherein the type includes a plurality of user groups and a plurality of time dimensions;
the data processing module 402 is configured to perform dimension reduction processing on the electricity sales volume sample data to obtain feature data corresponding to different user groups in different time dimensions;
the reference value calculation module 403 is configured to input the feature data into a preset prediction model library to obtain a sales power prediction reference value;
the parameter calculation module 404 is configured to input the feature data and the sales amount prediction reference value into an optimization model to obtain optimization parameters corresponding to the feature data of different user groups in different time dimensions;
the optimizing module 405 is configured to adjust a preset prediction model library according to the optimizing parameters to obtain an optimized prediction model library;
and the prediction module 406 is configured to input the feature data into the optimized prediction model library, and perform medium-long term prediction on the sales power under different types to obtain a medium-long term sales power predicted value of the power grid.
The parameter calculation module 404 may be further configured to:
calculating a first average value corresponding to the characteristic data and a second average value corresponding to the sales quantity prediction reference value;
obtaining a distance reference value based on the difference between the first average value and the second average value;
initializing the difference distance between two adjacent full-connection layers in the optimization model according to the distance reference value to obtain an initialized optimization model;
and obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the initialized optimization model.
The reference value calculation module 403 may be further configured to:
acquiring a first loss value; the first loss value is an error between the characteristic data and the sales quantity prediction reference value.
The parameter calculation module 404 may be further configured to:
inputting the characteristic data into the initialized optimization model to predict the sales amount;
calculating the difference distance between two adjacent full-connection layers in the initialized optimization model in the prediction process;
constructing an objective function according to the first loss value and the difference distance;
obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the minimum value of the objective function; wherein the optimization parameters include depth and width of the fully connected layer.
The data processing module 402 may also be configured to:
calculating a corresponding global divergence matrix according to the electricity sales volume sample data;
obtaining a projection matrix corresponding to the sales quantity sample data according to the global divergence matrix;
and obtaining characteristic data corresponding to different user groups under different time dimensions according to the projection matrix and the sales volume sample data.
The reference value calculation module 403 may be further configured to:
searching a prediction model matched with the feature type in a preset prediction model library based on the feature type of the feature data to obtain a reference sequence;
calculating the weight of each prediction model in the reference sequence, and determining the prediction model with the largest weight in the reference sequence as a reference model;
and inputting the characteristic data into a reference model to obtain a sales quantity prediction reference value.
The optimizing module 405 may be further configured to:
and adjusting a reference model in a preset prediction model library based on the optimization parameters to obtain an optimized reference model.
The prediction module 406 may be further configured to:
and inputting the characteristic data into the optimized reference model to obtain the medium-and-long-term sales quantity predicted value corresponding to the characteristic type of the characteristic data.
The embodiment of the application provides a medium-and-long-term electricity sales quantity prediction device, which is characterized in that an acquisition module is used for acquiring electricity sales quantity sample data under different types in a power grid, a data processing module is used for processing the acquired electricity sales quantity sample data to obtain characteristic data, a reference value calculation module is used for acquiring an electricity sales quantity prediction reference value, an optimization module is then used for adjusting a preset prediction model library by using optimization parameters acquired by a parameter calculation module, and finally the medium-and-long-term electricity sales quantity is predicted by the prediction module to acquire an electricity sales quantity prediction value. The embodiment of the application can process the sample data of the sales power quantity of different types and accurately predict the sales power quantity with different change rules.
Fig. 5 is a schematic diagram of a terminal according to an embodiment of the present application. As shown in fig. 5, the terminal 500 of this embodiment includes: a processor 501, a memory 502 and a computer program 503 stored in said memory 502 and executable on said processor 501. The processor 501, when executing the computer program 503, implements the steps of the embodiments of the medium-long term sales volume prediction method described above, such as steps 101 to 106 shown in fig. 1. Alternatively, the processor 501 may implement the functions of the modules/units in the above-described embodiments of the apparatus, such as the functions of the modules/units 401 to 406 shown in fig. 4, when executing the computer program 503.
Illustratively, the computer program 503 may be split into one or more modules/units that are stored in the memory 502 and executed by the processor 501 to accomplish the present application. The one or more modules/units may be a series of computer program instruction segments capable of performing the specified functions, which instruction segments describe the execution of the computer program 503 in the terminal 500. For example, the computer program 503 may be split into modules/units 401 to 406 shown in fig. 4.
The terminal 500 may include, but is not limited to, a processor 501, a memory 502. It will be appreciated by those skilled in the art that fig. 5 is merely an example of a terminal 500 and is not intended to limit the terminal 500, and may include more or fewer components than shown, or may combine certain components, or different components, e.g., the terminal may further include input and output devices, network access devices, buses, etc.
The processor 501 may be a central processing unit (Central Processing Unit, CPU), but may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The storage 502 may be an internal storage unit of the terminal 500, for example, a hard disk or a memory of the terminal 500. The memory 502 may also be an external storage device of the terminal 500, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash Card (Flash Card) or the like, which are provided on the terminal 500. Further, the memory 502 may also include both internal storage units and external storage devices of the terminal 500. The memory 502 is used for storing the computer program and other programs and data required by the terminal. The memory 502 may also be used to temporarily store data that has been output or is to be output.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions. The functional units and modules in the embodiment may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit, where the integrated units may be implemented in a form of hardware or a form of a software functional unit. In addition, the specific names of the functional units and modules are only for distinguishing from each other, and are not used for limiting the protection scope of the present application. The specific working process of the units and modules in the above system may refer to the corresponding process in the foregoing method embodiment, which is not described herein again.
In the foregoing embodiments, the descriptions of the embodiments are emphasized, and in part, not described or illustrated in any particular embodiment, reference is made to the related descriptions of other embodiments.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus/terminal and method may be implemented in other manners. For example, the apparatus/terminal embodiments described above are merely illustrative, e.g., the division of the modules or units is merely a logical function division, and there may be additional divisions when actually implemented, e.g., multiple units or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection via interfaces, devices or units, which may be in electrical, mechanical or other forms.
The units described as separate units may or may not be physically separate, and units shown as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated modules/units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the present application may implement all or part of the procedures in the methods of the above embodiments, or may be implemented by instructing the relevant hardware by a computer program, where the computer program may be stored in a computer readable storage medium, and the computer program may implement the steps of the embodiments of the method for predicting medium-to-long-term sales power when executed by a processor. Wherein the computer program comprises computer program code which may be in source code form, object code form, executable file or some intermediate form etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth.
The above embodiments are only for illustrating the technical solution of the present application, and not for limiting the same; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application, and are intended to be included in the scope of the present application.
Claims (10)
1. The medium-and-long-term electricity sales quantity prediction method is characterized by comprising the following steps of:
acquiring electricity sales quantity sample data under different types in a power grid; wherein the types include a plurality of user groups and a plurality of time dimensions;
performing dimension reduction processing on the electricity sales volume sample data to obtain feature data corresponding to different user groups in different time dimensions;
inputting the characteristic data into a preset prediction model library to obtain a sales quantity prediction reference value;
inputting the characteristic data and the sales quantity prediction reference value into an optimization model to obtain optimization parameters corresponding to the characteristic data of different user groups in different time dimensions;
adjusting a preset prediction model library according to the optimization parameters to obtain an optimized prediction model library;
and inputting the characteristic data into the optimized prediction model library, and performing medium-long-term prediction on the sales power under different types to obtain a medium-long-term sales power predicted value of the power grid.
2. The method for predicting the medium-and-long-term electricity sales amount according to claim 1, wherein the optimization model includes a plurality of fully connected layers, the inputting the feature data and the electricity sales amount prediction reference value into the optimization model to obtain optimization parameters corresponding to the feature data of different user groups in different time dimensions includes:
calculating a first average value corresponding to the characteristic data and a second average value corresponding to the sales quantity prediction reference value;
obtaining a distance reference value based on the difference between the first average value and the second average value;
initializing the difference distance between two adjacent full-connection layers in the optimization model according to the distance reference value to obtain an initialized optimization model;
and obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the initialized optimization model.
3. The method for predicting the long-term sales power according to claim 2, wherein after inputting the feature data into a preset prediction model library to obtain a sales power prediction reference value, the method further comprises:
acquiring a first loss value; the first loss value is an error between the characteristic data and the sales quantity prediction reference value;
and obtaining optimization parameters corresponding to the feature data of different user groups in different time dimensions according to the initialized optimization model, wherein the optimization parameters comprise:
inputting the characteristic data into the initialized optimization model to predict the sales amount;
calculating the difference distance between two adjacent full-connection layers in the initialized optimization model in the prediction process;
constructing an objective function according to the first loss value and the difference distance;
obtaining optimization parameters corresponding to the characteristic data of different user groups under different time dimensions according to the minimum value of the objective function; wherein the optimization parameters include depth and width of the full link layer.
4. The method for predicting medium-to-long-term electricity sales volume according to claim 1, wherein the performing dimension reduction processing on the electricity sales volume sample data to obtain feature data corresponding to different user groups in different time dimensions comprises:
calculating a corresponding global divergence matrix according to the electricity sales volume sample data;
obtaining a projection matrix corresponding to the electricity sales quantity sample data according to the global divergence matrix;
and obtaining characteristic data corresponding to different user groups under different time dimensions according to the projection matrix and the sales volume sample data.
5. The method of claim 1, wherein the pre-set prediction model library comprises a plurality of different prediction models.
6. The method for predicting the medium-and-long-term electricity sales amount according to claim 5, wherein each prediction model is provided with a tag, the type of the tag corresponds to the type of the electricity sales amount sample data, the inputting the feature data into a preset prediction model library to obtain an electricity sales amount prediction reference value includes:
searching a prediction model matched with the feature type in the preset prediction model library based on the feature type of the feature data to obtain a reference sequence; wherein the feature type corresponds to the type of the tag;
calculating the weight of each prediction model in the reference sequence, and determining the prediction model with the largest weight in the reference sequence as a reference model;
and inputting the characteristic data into the reference model to obtain a sales quantity prediction reference value.
7. The method for predicting medium-to-long-term electricity sales according to claim 6, wherein the adjusting the preset prediction model library according to the optimization parameters to obtain the optimized prediction model library comprises:
adjusting a reference model in the preset prediction model library based on the optimization parameters to obtain an optimized reference model;
inputting the characteristic data into the optimized prediction model library, and performing medium-long term prediction on the sales power under different types to obtain a medium-long term sales power predicted value, wherein the method comprises the following steps:
and inputting the characteristic data into the optimized reference model to obtain a medium-and-long-term sales quantity predicted value corresponding to the characteristic type of the characteristic data.
8. A medium-to-long-term sales power prediction apparatus, comprising:
the acquisition module is used for acquiring the electricity sales quantity sample data under different types in the power grid; wherein the types include a plurality of user groups and a plurality of time dimensions;
the data processing module is used for performing dimension reduction processing on the electricity sales volume sample data to obtain characteristic data corresponding to different user groups in different time dimensions;
the reference value calculation module is used for inputting the characteristic data into a preset prediction model library to obtain a sales quantity prediction reference value;
the parameter calculation module is used for inputting the characteristic data and the sales quantity prediction reference value into an optimization model to obtain optimization parameters corresponding to the characteristic data of different user groups under different time dimensions;
the optimization module is used for adjusting a preset prediction model library according to the optimization parameters to obtain an optimized prediction model library;
and the prediction module is used for inputting the characteristic data into the optimized prediction model library, and performing medium-long-term prediction on the sales power under different types to obtain a medium-long-term sales power predicted value of the power grid.
9. A terminal comprising a memory for storing a computer program and a processor for calling and running the computer program stored in the memory, characterized in that the processor implements the steps of the medium-and-long-term sales volume prediction method according to any one of the preceding claims 1 to 7 when the computer program is executed.
10. A computer readable storage medium storing a computer program, characterized in that the computer program when executed by a processor implements the steps of the medium-to-long term sales volume prediction method of any one of the preceding claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310797117.6A CN116822739A (en) | 2023-06-30 | 2023-06-30 | Medium-long-term sales electricity quantity prediction method, device, terminal and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310797117.6A CN116822739A (en) | 2023-06-30 | 2023-06-30 | Medium-long-term sales electricity quantity prediction method, device, terminal and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116822739A true CN116822739A (en) | 2023-09-29 |
Family
ID=88127244
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310797117.6A Pending CN116822739A (en) | 2023-06-30 | 2023-06-30 | Medium-long-term sales electricity quantity prediction method, device, terminal and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116822739A (en) |
-
2023
- 2023-06-30 CN CN202310797117.6A patent/CN116822739A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113239314A (en) | Method, device, terminal and computer-readable storage medium for carbon emission prediction | |
| CN111091196B (en) | Passenger flow data determination method and device, computer equipment and storage medium | |
| CN106549772A (en) | Resource prediction method, system and capacity management device | |
| CN112990587B (en) | Method, system, equipment and medium for accurately predicting power consumption of transformer area | |
| CN113255973A (en) | Power load prediction method, power load prediction device, computer equipment and storage medium | |
| CN109741177A (en) | Appraisal procedure, device and the intelligent terminal of user credit | |
| CN109726858A (en) | Heat load prediction method and device based on dynamic time warping | |
| CN113283936A (en) | Sales forecasting method, sales forecasting device and electronic equipment | |
| CN116187835A (en) | Data-driven-based method and system for estimating theoretical line loss interval of transformer area | |
| CN116579804A (en) | Holiday commodity sales prediction method, holiday commodity sales prediction device and computer storage medium | |
| CN117196695B (en) | Target product sales data prediction method and device | |
| CN110309947A (en) | Complete vehicle logistics order forecast method and device, logistics system and computer-readable medium | |
| CN113902260A (en) | Information prediction method, information prediction device, electronic equipment and medium | |
| CN116245259B (en) | Photovoltaic power generation prediction method and device based on depth feature selection and electronic equipment | |
| CN109800138B (en) | CPU testing method, electronic device and storage medium | |
| CN116822739A (en) | Medium-long-term sales electricity quantity prediction method, device, terminal and storage medium | |
| CN116191398A (en) | Load prediction method, load prediction device, computer equipment and storage medium | |
| CN112926801B (en) | Load curve combined prediction method and device based on quantile regression | |
| CN115034812A (en) | Steel industry sales prediction method and device based on big data | |
| CN115544860A (en) | Output modeling method of intermittent distributed power supply in complex operation scene | |
| CN111768282B (en) | Data analysis method, device, equipment and storage medium | |
| CN114970357A (en) | Energy-saving effect evaluation method, system, device and storage medium | |
| CN114549233A (en) | Floating population prediction method based on combination of LGB algorithm and ARIMA algorithm | |
| CN117435870B (en) | Load data real-time filling method, system, equipment and medium | |
| CN116562359B (en) | CTR prediction model training method and device based on contrast learning and electronic equipment |
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 |