CN116862053A - Load curve prediction method and device based on hybrid model - Google Patents

Load curve prediction method and device based on hybrid model Download PDF

Info

Publication number
CN116862053A
CN116862053A CN202310699170.2A CN202310699170A CN116862053A CN 116862053 A CN116862053 A CN 116862053A CN 202310699170 A CN202310699170 A CN 202310699170A CN 116862053 A CN116862053 A CN 116862053A
Authority
CN
China
Prior art keywords
prediction
data
load
load curve
annual maximum
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310699170.2A
Other languages
Chinese (zh)
Inventor
姚知洋
李珊
黄维
周杨珺
欧阳健娜
陆新
张炜
刘鹏
张玉波
奉斌
唐捷
黎蓓
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Electric Power Research Institute of Guangxi Power Grid Co Ltd
Original Assignee
Electric Power Research Institute of Guangxi Power Grid Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Electric Power Research Institute of Guangxi Power Grid Co Ltd filed Critical Electric Power Research Institute of Guangxi Power Grid Co Ltd
Priority to CN202310699170.2A priority Critical patent/CN116862053A/en
Publication of CN116862053A publication Critical patent/CN116862053A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • 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
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/003—Load forecast, e.g. methods or systems for forecasting future load demand

Abstract

The application discloses a load curve prediction method and device based on a hybrid model, belonging to the technical field of load prediction, wherein the method comprises the following steps: carrying out load curve trend prediction through a load curve prediction model to obtain load curve prediction data; carrying out annual maximum load data prediction through an annual maximum load prediction model to obtain an annual maximum load data prediction value; and carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result. Therefore, the load curve prediction is realized, and the problem of inaccurate medium-year load prediction is solved.

Description

Load curve prediction method and device based on hybrid model
Technical Field
The application relates to the technical field of load prediction, in particular to a load curve prediction method and device based on a hybrid model.
Background
The load curve prediction is an extremely important part of the electric power market, the prediction result of the load curve provides decision basis for the electric power installed capacity and the resource scheduling, and the accuracy of the prediction directly determines the power grid layout, investment and scheduling compliance and rationality. With the continued development of the power market, conventional load curve prediction levels have failed to meet the increasing demands of the power market. In addition, along with the development of socioeconomic performance, the factors influencing the planning, running and scheduling of the power grid are more and more, and the prediction accuracy of the load curve is difficult to ensure by a single prediction method.
The existing patent A, namely a load weighted integration prediction method based on differentiated multimode fusion, provides a load weighted integration prediction method based on linear regression, a Markov transfer matrix and a BP neural network; the second patent provides a support vector machine-based method for predicting a minute-scale load curve; the third patent refers to a baseline load prediction correction method based on user electricity consumption characteristics, which comprises the steps of classifying user characteristics of historical user electricity consumption load data by using K-MEANS, and correcting predicted user baseline load according to classification results; the patent IV is a load prediction method and flow based on superposition of multi-industry history typical load curves, which respectively predicts 8760 loads of each industry, and a prediction result is formed after superposition; the patent five provides a DCAE-LSTM-based short-term load curve prediction method, which utilizes a neural network to fit the trend of a load curve well.
The weighted integration method of the first patent can reduce the deviation between different methods, but cannot reduce the prediction error of the method itself, and is not suitable for the prediction of a large amount of data due to linear regression, and the speed is very slow; the support vector machine of the second patent is only suitable for short-term load prediction, is more suitable for small sample data, and has no advantage on a large amount of sample data; the three patents pass through the classification of the users to use electricity selling types or the characteristics of the industries where the users are located, and are not required to be processed by K-MEANS; the fourth and fifth patents have difficulty in capturing the maximum or minimum value of the annual load.
Disclosure of Invention
The embodiment of the application provides a load curve prediction method and device based on a hybrid model, which at least solve the technical problem of low load curve prediction accuracy in the related art.
According to an aspect of the embodiment of the present application, there is provided a load curve prediction method based on a hybrid model, including:
acquiring data required by load prediction, cleaning and preprocessing the data required by the load prediction, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting a load curve trend through the load curve prediction model to obtain load curve prediction data;
acquiring data required by annual maximum load prediction, cleaning and preprocessing the data required by the annual maximum load prediction, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value;
and carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result.
Optionally, the data required for load prediction includes: historical time characteristic data, historical weather data, historical holiday data and historical load curve data.
Optionally, preprocessing the data required for load prediction includes: extracting characteristic values from the historical time characteristic data; converting the historical weather data into numerical data; coding the historical holiday data; and carrying out normalization processing on all data required by load prediction and then carrying out data division.
Optionally, the load curve prediction model employs LSTM for time series data prediction.
Optionally, the data required for annual maximum load prediction includes: time characteristic data, economic data, and historical maximum load data.
Optionally, performing reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result, including:
obtaining the maximum load of a load curve from load curve prediction data;
carrying out standardization processing on the load curve prediction data by combining the maximum load of the load curve;
and multiplying the annual maximum load data predicted value by the normalized processing result to obtain a final load curve predicted result.
Optionally, the normalization process divides each value in the load curve prediction data by the load curve maximum load.
According to another aspect of the embodiment of the present application, there is also provided a load curve prediction apparatus based on a hybrid model, including:
the load curve prediction module is used for acquiring data required by load prediction, cleaning and preprocessing the data required by the load prediction, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting a load curve trend through the load curve prediction model to obtain load curve prediction data;
the annual maximum load data prediction module is used for acquiring data required by annual maximum load prediction, cleaning and preprocessing the data required by the annual maximum load prediction, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value;
and the reverse correction module is used for carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result.
According to another aspect of the embodiment of the present application, there is also provided a load curve prediction system based on a hybrid model, including: one or more processors; a storage means for storing one or more programs; the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the hybrid model-based load curve prediction method of any of the above.
According to another aspect of the embodiment of the present application, there is further provided a computer readable storage medium, where the computer readable storage medium includes a stored program, and when the program runs, controls a device in which the computer readable storage medium is located to execute the load curve prediction method based on the hybrid model according to any one of the above.
Compared with the prior art, the application has the following beneficial effects:
in the embodiment of the application, the advantages of strong linear or nonlinear fitting capacity of the neural network and the advantages of accurate long-term load prediction of the traditional load prediction are combined, the trend prediction is carried out on the load curve by using the neural network, and then the load curve is reversely corrected by the traditional load prediction result. Load curve apertures include, but are not limited to, full social aperture, dispatch aperture, transaction aperture, and marketing aperture. The load curve prediction model established by the application relates to provincial total load curve and load curve prediction of the grading market, and the accuracy of the load curve prediction model can reach about 98%.
Drawings
In order to more clearly illustrate the technical solutions of the present application, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawing in the description below is only one embodiment 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 a load curve prediction method based on a hybrid model according to an embodiment of the present application.
Detailed Description
It should be noted that, without conflict, the embodiments of the present application and features of the embodiments may be combined with each other. The application will be described in detail below with reference to the drawings in connection with embodiments.
In order that those skilled in the art will better understand the present application, a technical solution in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present application without making any inventive effort, shall fall within the scope of the present application.
It should be noted that the terms "first," "second," and the like in the description and the claims of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate in order to describe the embodiments of the application herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, apparatus, article, or device that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or device.
Example 1
According to an embodiment of the present application, there is provided an embodiment of a load curve prediction method based on a hybrid model, it should be noted that the steps illustrated in the flowchart of the drawings may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is illustrated in the flowchart, in some cases the steps illustrated or described may be performed in an order different from that herein.
FIG. 1 is a flowchart of a load curve prediction method based on a hybrid model according to an embodiment of the present application, and as shown in FIG. 1, the load curve prediction includes the steps of:
step S1, acquiring data required by load prediction, cleaning and preprocessing the data required by the load prediction, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting a load curve trend through the load curve prediction model to obtain load curve prediction data;
s2, acquiring data required by annual maximum load prediction, cleaning and preprocessing the data required by annual maximum load prediction, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value;
and S3, carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result.
As an alternative embodiment, step S1 comprises the steps of:
step S11, acquiring data required by load prediction, wherein the data required by the load prediction comprises the following steps: historical time characteristic data, historical weather data, historical holiday data and historical load curve data.
Specifically, the minimum frequency of the historical time feature data is hour; the minimum frequency of historical weather data is the day; the historical holiday data is directly marked on the historical time characteristic data, and the minimum frequency is the day; the minimum frequency of the historical load curve data is hours.
And step S12, cleaning data required by load prediction so as to meet the use requirement. The data cleaning rule comprises the following steps:
(1) Deleting the acquired null value;
(2) Performing deduplication on the acquired and repeated data;
(3) Deleting the data with the wrong range, wherein the data ranges of different labels are different;
(4) The data acquired through the weather interface is purged, for example, data that appears to be outside the weather type range, and the reliability of the historical load data is verified.
Step S13, preprocessing the data required by load prediction after the cleaning in step S12, specifically comprising the following steps:
step S131, extracting characteristic values from the historical time characteristic data.
Specifically, sequentially extracting the time features includes: five characteristic values of year, season, month, day and time, namely characteristic value A 1 ＝[a 1 ,a 2 ,a 3 ,a 4 ,a 5 ]，a 1 ,a 2 ,a 3 ,a 4 ,a 5 Respectively year, season, month, day and time. For example, 2023, 1.1.1.1.1.points, and extracting the characteristic value to obtain [2023,1,1,1,1 ]]。
Step S132, converting the historical weather data into numerical data.
Specifically, few models can be directly modeled by using text data, data coding is carried out on different types of weather data, and the weather data is converted into a digital type by adopting a single-hot code coding mode.
And step S133, encoding the historical holiday data.
Specifically, holiday data related to the present application includes legal holidays, saturday, region-specific holidays, such as Dai's splatter water festival, etc. specified by the country. Considering the influence of different holidays and holiday lengths on load, the application adopts different weight codes for the holidays, and the specific coding mode is to code according to the holiday releasing days of the holidays, for example, the working day codes are all 0, the Saturday codes are [1,2], and the labor sections are all five days and the codes are [1,2,3,4,5]. Considering that the holiday minimum granularity is day and the load curve data minimum granularity is hour, the holiday code is repeated 24 times to dimensionally match the load curve data.
Step S134, normalizing the data required by load prediction.
Specifically, the data normalization can eliminate the trouble brought by the data dimension, and is particularly in the neural network, the data normalization is favorable for network initialization, the numerical problem brought by updating the gradient numerical value can be avoided, and the adjustment of the learning rate numerical value and the speed of searching the optimal solution are also favorable. Common data normalization methods comprise data normalization, data maximization and the like.
And step S134, carrying out data proportion division on the data obtained by processing in the step S134 to obtain training data, test data and verification data.
Specifically, the application adopts the method that the data is segmented into training data, test data and verification data according to the proportion of 6:2:2.
And S14, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting the trend of the load curve through the load curve prediction model to obtain the load curve prediction data. The method specifically comprises the following steps:
s141, constructing a model, wherein a load curve prediction model is constructed by load prediction data, and the LSTM model is constructed by tensorflow;
s142, training the model, namely training the model obtained in the step S141, inputting training data and verification data into a load curve prediction model, and training the load curve prediction model;
s143, evaluating the model, namely evaluating the load curve prediction model obtained by training in the step S142, and finishing the evaluation by adopting scores of training data, test data and verification data on the load curve prediction model;
s144, storing the model, namely storing the load curve prediction model into a model file after the steps S411-S413 so as to be convenient for subsequent calling.
The application of the load curve prediction model needs to consider two problems, namely that the constructed model meets the requirements, and a new characteristic value needs to be transmitted into the model. The new input characteristic values include future time characteristic data, encoded future weather data, encoded future holiday data.
Among the most difficult to obtain is future weather data, which needs to be crawled from a weather website or obtained through an interface, and the maximum time period of the acquisition is not more than one year. And (3) preprocessing and normalizing the prepared data, and inputting the preprocessed and normalized data into a model to obtain the load curve prediction data of the corresponding time period.
As an alternative embodiment, the economic development and the electric power development are complementary, the electric power is a key basis of the economic development, and the economic development also promotes the electric power development. In particular, in recent years, policies are in progress, economy in each region is rapidly developed, sales power is continuously increased, load demands are continuously increased, and unit capacity of power generation enterprises is continuously increased. The application utilizes the relation between economy and electricity to complete the prediction of annual maximum load. The step S2 specifically comprises the following steps:
s21, acquiring data required by annual maximum load prediction, wherein the data required by annual maximum load prediction comprises: time characteristic data, economic data (specifically, domestic total production value, which has strong correlation with electric quantity or load) and historical maximum load data.
Specifically, the temporal feature data adopts an annual temporal feature. Each place has professional statistical websites to publish regional annual economy data, so regional historical economy data can be obtained from the websites, and the regional historical economy data required by the application comprises: the first industry nominal added value, the second industry nominal added value, and the third industry nominal added value are each year in frequency. The frequency of the historical maximum load data is annual, and the maximum load data aperture herein includes, but is not limited to, a full social aperture, a dispatch aperture, a trade aperture, and a marketing aperture.
S22, cleaning data required by annual maximum load prediction; the data cleaning rule comprises the following steps:
(1) Deleting the acquired null value;
(2) Performing deduplication on the acquired and repeated data;
(3) Deleting the data with the wrong range, wherein the data ranges of different labels are different;
(4) And cleaning the data acquired through the economic interface and verifying the reliability of the historical load data.
S23, preprocessing the data required by the annual maximum load prediction after the cleaning in the step S13.
Specifically, the historical data related to the annual maximum load data prediction are all digital data, so that the historical maximum load data only needs to be normalized.
S24, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value. The method specifically comprises the following steps:
s241, constructing a model, wherein the annual maximum load prediction model adopts linear regression modeling, and specifically, adopting LinearRegression, ridge or Lasso in sklearn to complete the annual maximum load prediction model modeling;
s242, training the model, namely training the model obtained in the step S241, inputting training data and verification data into the annual maximum load prediction model, and training the annual maximum load prediction model;
s243, evaluating the model, namely evaluating the annual maximum load prediction model obtained by training in the step S242, and finishing the evaluation by adopting scores of training data, test data and verification data on the annual maximum load prediction model;
s244, saving the model, namely saving the annual maximum load prediction model into a model file after the steps S421-S423 so as to be convenient for subsequent calling.
As an optional embodiment, step S3 of reversely correcting the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result, includes the following steps:
s31, obtaining the maximum load of a load curve from load curve prediction data;
s32, carrying out standardization processing on the load curve prediction data by combining the maximum load of the load curve; wherein the normalization process divides each value in the load curve prediction data by the load curve maximum load;
and step S33, multiplying the annual maximum load data predicted value by the normalized processing result to obtain a final load curve predicted result.
Specifically, the load curve is reversely corrected through the annual maximum load predicted value, so that the trend value of the load curve meets the predicted requirement. Taking 8760 load curve correction for the future year as an example, 8760 load curve is set as a List [ a ] 0 ,a 1 ,...,a 8759 ]The correction procedure is as follows (List [ a ] 0 ,a 1 ,...,a 8759 ]For the prediction results, l_max is the annual maximum load prediction result, which is known at the time of calculation here), is also known at the time of calculation here):
step S31, obtaining a load curve maximum load from load curve prediction data, and setting the load curve maximum load as L_max, wherein:
L_max＝max(List[a 0 ,a 1 ,...,a 8759 ]) Where max is the maximum calculation function.
Step S32, normalizing the annual data list of the load curve
Let the List of the normalized annual data of the load curve be List [ b ] 0 ,b 1 ,...,b 8759 ]Then:
List[b 0 ,b 1 ,...,b 8759 ]＝List[a 0 ,a 1 ,...,a 8759 ]l_max, represents List [ a ] 0 ,a 1 ,...,a 8759 ]Divided by L _ max.
Step S33, load curve reverse correction
The maximum load annual prediction value is set as l_max, which can be obtained from the result of annual maximum load prediction in step S2 of the present application, and the modified load curve annual data is set as List [ c ] 0 ,c 1 ,...,c 8759 ]Then:
List[c 0 ,c 1 ,...,c 8759 ]＝List[b 0 ,b 1 ,...,b 8759 ]* l_max, represents List [ b ] 0 ,b 1 ,...,b 8759 ]Is multiplied by l _ max. And finally, finishing the calculation.
Example 2
According to another aspect of the embodiment of the present application, there is also provided a load curve prediction apparatus based on a hybrid model, the load curve prediction apparatus based on the hybrid model applying the above load curve prediction method based on the hybrid model, the load curve prediction apparatus including:
the load curve prediction module is used for acquiring data required by load prediction, cleaning and preprocessing the data required by the load prediction, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting a load curve trend through the load curve prediction model to obtain load curve prediction data;
the annual maximum load data prediction module is used for acquiring data required by annual maximum load prediction, cleaning and preprocessing the data required by annual maximum load prediction, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value;
and the reverse correction module is used for carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result.
The present application is not limited to the above embodiments, but is to be accorded the widest scope consistent with the principles and other features of the present application.
Example 3
According to another aspect of the embodiment of the present application, there is also provided a load curve prediction system based on a hybrid model, including: one or more processors; a storage means for storing one or more programs; the one or more programs, when executed by the one or more processors, cause the one or more processors to implement the hybrid model-based load curve prediction method of any of the above.
Example 4
According to another aspect of the embodiments of the present application, there is further provided a computer readable storage medium including a stored program, wherein the apparatus in which the computer readable storage medium is controlled to execute the hybrid model-based load curve prediction method of any one of the above when the program runs.
Alternatively, in this embodiment, the above-mentioned computer readable storage medium may be located in any one of the computer terminals in the computer terminal group in the computer network or in any one of the mobile terminals in the mobile terminal group, and the above-mentioned computer readable storage medium includes a stored program.
Optionally, the computer readable storage medium is controlled to perform the following functions when the program is run: acquiring data required by load prediction, cleaning and preprocessing the data required by the load prediction, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting a load curve trend through the load curve prediction model to obtain load curve prediction data;
acquiring data required by annual maximum load prediction, cleaning and preprocessing the data required by the annual maximum load prediction, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value;
and carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result.
The foregoing embodiment numbers of the present application are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
In the foregoing embodiments of the present application, the descriptions of the embodiments are emphasized, and for a portion of this disclosure that is not described in detail in this embodiment, reference is made to the related descriptions of other embodiments.
In the several embodiments provided in the present application, it should be understood that the disclosed technology may be implemented in other manners. The above-described embodiments of the apparatus are merely exemplary, and the division of the units, for example, may be a logic function division, and may be implemented in another manner, for example, a plurality of units or components may be combined or may be integrated into another apparatus, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some interfaces and the indirect coupling or communication connection of units or modules may be in electrical or other forms.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of 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 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 technical solution of the present application may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a Read-0nlyMemory (ROM), a random access memory (RAM, randomAccessMemory), a removable hard disk, a magnetic disk, or an optical disk, or the like, which can store program codes.
The foregoing is merely a preferred embodiment of the present application and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present application, which are intended to be comprehended within the scope of the present application.

Claims (10)

1. A load curve prediction method based on a hybrid model, comprising:
acquiring data required by load prediction, cleaning and preprocessing the data required by the load prediction, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting a load curve trend through the load curve prediction model to obtain load curve prediction data;
acquiring data required by annual maximum load prediction, cleaning and preprocessing the data required by the annual maximum load prediction, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value;
and carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result.
2. The hybrid model-based load curve prediction method according to claim 1, wherein the data required for the load prediction includes: historical time characteristic data, historical weather data, historical holiday data and historical load curve data.
3. The hybrid model based load curve prediction method of claim 1, wherein preprocessing the data required for load prediction comprises: extracting characteristic values from the historical time characteristic data; converting the historical weather data into numerical data; coding the historical holiday data; and carrying out normalization processing on all data required by load prediction and then carrying out data division.
4. The hybrid model-based load curve prediction method of claim 1, wherein the load curve prediction model employs LSTM for time series data prediction.
5. The hybrid model based load curve prediction method according to claim 1, wherein the data required for annual maximum load prediction includes: time characteristic data, economic data, and historical maximum load data.
6. The load curve prediction method based on the hybrid model according to claim 1, wherein the performing inverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result includes:
obtaining the maximum load of a load curve from load curve prediction data;
carrying out standardization processing on the load curve prediction data by combining the maximum load of the load curve;
and multiplying the annual maximum load data predicted value by the normalized processing result to obtain a final load curve predicted result.
7. The hybrid model based load curve prediction method of claim 6, wherein the normalization process is to divide each value in the load curve prediction data by a load curve maximum load.
8. A hybrid model-based load curve prediction apparatus, comprising:
the load curve prediction module is used for acquiring data required by load prediction, cleaning and preprocessing the data required by the load prediction, establishing a load curve prediction model according to the preprocessed load prediction data, and predicting a load curve trend through the load curve prediction model to obtain load curve prediction data;
the annual maximum load data prediction module is used for acquiring data required by annual maximum load prediction, cleaning and preprocessing the data required by the annual maximum load prediction, establishing an annual maximum load prediction model according to the preprocessed annual maximum load prediction data, and predicting the annual maximum load data through the annual maximum load prediction model to obtain an annual maximum load data prediction value;
and the reverse correction module is used for carrying out reverse correction on the load curve prediction data according to the annual maximum load data prediction value to obtain a final load curve prediction result.
9. A hybrid model-based load curve prediction system, the hybrid model-based load curve prediction system comprising:
one or more processors;
a storage means for storing one or more programs;
the one or more programs, when executed by one or more processors, cause the one or more processors to implement the hybrid model-based load curve prediction method of any of claims 1 to 7.
10. A computer readable storage medium, characterized in that the computer readable storage medium comprises a stored program, wherein the program when run controls a device in which the computer readable storage medium is located to perform the hybrid model based load curve prediction method of any one of claims 1 to 7.
CN202310699170.2A 2023-06-13 2023-06-13 Load curve prediction method and device based on hybrid model Pending CN116862053A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310699170.2A CN116862053A (en) 2023-06-13 2023-06-13 Load curve prediction method and device based on hybrid model

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310699170.2A CN116862053A (en) 2023-06-13 2023-06-13 Load curve prediction method and device based on hybrid model

Publications (1)

Publication Number Publication Date
CN116862053A true CN116862053A (en) 2023-10-10

Family

ID=88225852

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310699170.2A Pending CN116862053A (en) 2023-06-13 2023-06-13 Load curve prediction method and device based on hybrid model

Country Status (1)

Country Link
CN (1) CN116862053A (en)

Similar Documents

Publication Publication Date Title
CN110400022B (en) Cash consumption prediction method and device for self-service teller machine
WO2019056499A1 (en) Prediction model training method, data monitoring method, apparatuses, device and medium
CN110689190A (en) Power grid load prediction method and device and related equipment
Alamsyah et al. Artificial neural network for Indonesian tourism demand forecasting
CN111340375A (en) Electricity charge recycling risk prediction method and device, electronic equipment and storage medium
CN110147389A (en) Account number treating method and apparatus, storage medium and electronic device
CN112614011A (en) Power distribution network material demand prediction method and device, storage medium and electronic equipment
CN112418476A (en) Ultra-short-term power load prediction method
CN111951045A (en) Electricity charge accounting method and device based on big data
CN111178957B (en) Method for early warning sudden increase of electric quantity of electricity consumption customer
CN114943565A (en) Electric power spot price prediction method and device based on intelligent algorithm
CN116915710A (en) Traffic early warning method, device, equipment and readable storage medium
CN114584601A (en) User loss identification and intervention method, system, terminal and medium
CN114331001A (en) Self-adaptive matching method and system for allocating abnormal work orders of power terminal
CN116523301A (en) System for predicting risk rating based on big data of electronic commerce
CN116862053A (en) Load curve prediction method and device based on hybrid model
CN114282657A (en) Market data long-term prediction model training method, device, equipment and storage medium
CN115238940A (en) Training method and device of power grid load prediction model
CN114140176A (en) Adjustable capacity prediction method and device for load aggregation platform
CN113506173A (en) Credit risk assessment method and related equipment thereof
CN112614006A (en) Load prediction method, device, computer readable storage medium and processor
CN116562398B (en) Power load prediction model training method, prediction method, electronic device and medium
CN116011698B (en) Method, device, computer equipment and storage medium for determining unit combination
CN117236532B (en) Load data-based electricity consumption peak load prediction method and system
CN113837486B (en) RNN-RBM-based distribution network feeder long-term load prediction method

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