CN113255973A - Power load prediction method, power load prediction device, computer equipment and storage medium - Google Patents
Power load prediction method, power load prediction device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN113255973A CN113255973A CN202110503715.9A CN202110503715A CN113255973A CN 113255973 A CN113255973 A CN 113255973A CN 202110503715 A CN202110503715 A CN 202110503715A CN 113255973 A CN113255973 A CN 113255973A
- Authority
- CN
- China
- Prior art keywords
- load
- prediction model
- time period
- load prediction
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000003860 storage Methods 0.000 title claims abstract description 18
- 238000012360 testing method Methods 0.000 claims description 55
- 238000012545 processing Methods 0.000 claims description 33
- 238000007781 pre-processing Methods 0.000 claims description 14
- 238000004590 computer program Methods 0.000 claims description 13
- 238000010606 normalization Methods 0.000 claims description 13
- 238000005457 optimization Methods 0.000 claims description 12
- 238000012549 training Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 claims description 11
- 238000004458 analytical method Methods 0.000 claims description 9
- 238000000605 extraction Methods 0.000 claims description 8
- 230000000737 periodic effect Effects 0.000 abstract description 7
- 238000004422 calculation algorithm Methods 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 13
- 230000000694 effects Effects 0.000 description 8
- 230000008859 change Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 238000004891 communication Methods 0.000 description 5
- 238000012706 support-vector machine Methods 0.000 description 5
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 4
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 4
- 235000019580 granularity Nutrition 0.000 description 4
- 238000010248 power generation Methods 0.000 description 4
- 238000001556 precipitation Methods 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 230000002354 daily effect Effects 0.000 description 2
- 230000003203 everyday effect Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000013139 quantization Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000012512 characterization method Methods 0.000 description 1
- 230000002301 combined effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000001816 cooling Methods 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000010438 heat treatment Methods 0.000 description 1
- 238000003973 irrigation Methods 0.000 description 1
- 230000002262 irrigation Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Economics (AREA)
- Software Systems (AREA)
- Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Tourism & Hospitality (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Primary Health Care (AREA)
- General Health & Medical Sciences (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Development Economics (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The application relates to a power load prediction method, a power load prediction device, a computer device and a storage medium. The method comprises the following steps: the computer equipment obtains historical power load data, divides the historical power load data into a plurality of load data sets according to a plurality of preset time periods, determines a corresponding target load prediction model according to the time period of each load data set, and inputs each load data set into the corresponding target load prediction model to obtain a power load prediction result of the time period. According to the scheme, due to the fact that the power load data in different time periods have large differences, the target load prediction models corresponding to the time periods are built for predicting the power loads, compared with a traditional algorithm model, the prediction models in different time periods can better cope with the influences of periodic variation fluctuation and multi-factors of the power loads in different time periods, the method has high practicability, and the accuracy of the power load prediction results in the time periods is improved.
Description
Technical Field
The present application relates to the field of power technologies, and in particular, to a power load prediction method, apparatus, computer device, and storage medium.
Background
With the rapid development of economy, the demand of society for electric power resources has also been continuously increased. Because the electric energy is difficult to store in a large scale, the power load prediction has important significance for power generation and scheduling planning of power supply enterprises. In recent years, the scales of power grids and power loads are rapidly increasing, and higher requirements are made on the accuracy of Short-term load forecasting (STLF) and the stability of a forecasting model.
At present, the prediction of the short-term power load can be realized by constructing a prediction model, wherein the prediction model can be a fuzzy logic model (fuzzy logic system), a Support Vector Machine (SVM), an Artificial Neural Network (ANN), a Recurrent Neural Network (RNN), and the like, and the power load prediction is realized through the prediction model.
However, the above-described prediction of the power load by constructing the neural network prediction model has a problem of low prediction accuracy.
Disclosure of Invention
In view of the above, it is necessary to provide a power load prediction method, a power load prediction apparatus, a computer device, and a storage medium, which can improve the prediction accuracy.
In a first aspect, a method for predicting a power load is provided, the method comprising:
acquiring historical power load data;
dividing historical power load data into a plurality of load data sets according to a plurality of preset time periods;
determining a corresponding target load prediction model according to the time period of each load data set;
and inputting each load data set into a corresponding target load prediction model to obtain a power load prediction result of a time period.
In the embodiment, since the power load data of different time periods have large differences, in order to predict the power load of each time period in a targeted manner, the target load prediction model corresponding to each time period is constructed to predict the power load, so that the accuracy of the power load prediction result of each time period is improved.
In one embodiment, the target load prediction model includes at least two types of prediction models.
In this embodiment, the target load prediction model adopts a combined model, which can better cope with the influence of the periodic variation fluctuation and the multifactor of the power load in different time periods, and has strong practicability.
In one embodiment, the method for constructing the target load prediction model includes:
determining a plurality of preset time periods, and acquiring a sample load data set of each preset time period;
and training the initial load prediction model according to the sample load data set of each preset time period to obtain a target load prediction model corresponding to each preset time period.
In this embodiment, the initial power load prediction model of each time segment is trained specifically to obtain the target load prediction model of each time segment, and the prediction models of the time segments can improve the accuracy of the power load prediction result of each time segment.
In one embodiment, the training the initial load prediction model according to the sample load data set of each preset time period to obtain the target load prediction model corresponding to each preset time period includes:
for each sample load data set of a preset time period, dividing the sample load data set into first sample load data corresponding to a first sub-time period and second sample load data corresponding to a second sub-time period;
inputting the first sample load data into an initial load prediction model to obtain a prediction result;
and adjusting parameters of the initial load prediction model according to the prediction result and the second sample load data to obtain a target load prediction model.
In the embodiment, the data of the prediction time period in the virtual prediction is known, and the accuracy of the prediction model can be checked, so that the accuracy of the real prediction is improved.
In one embodiment, the initial load prediction model includes at least two types of prediction models, and the method further includes:
calculating the residual square sum of the prediction result and the second sample load data according to the prediction result output by the initial load prediction model of each preset time period;
and determining the weight of the initial load prediction model after the parameters are adjusted according to the residual sum of squares and a preset optimization objective function to obtain a target load prediction model.
In this embodiment, the computer device may optimize and solve to obtain the optimal weight of each model in the initial load prediction model by setting an optimized objective function and calculating the sum of squares of residuals of the prediction result and the second sample load data, so as to obtain the target load prediction model, and an output result of the optimized and solved target load prediction model is more accurate.
In one embodiment, the determining a plurality of preset time periods and obtaining the sample load dataset of each preset time period includes:
acquiring an original load data set of each preset time period;
carrying out data preprocessing on the original load data set to obtain a sample load data set of each preset time period;
the data preprocessing comprises at least one of data feature extraction, data missing value processing, data normalization processing, numeralization processing and data feature and power load association degree analysis.
In this embodiment, data feature extraction, data missing value processing, data normalization processing, digitization processing, and data feature and power load association degree analysis are performed on the original load data set, so as to obtain input data of the load prediction model under the normalization processing, and thus the obtained output result of the load prediction model is more accurate.
In one embodiment, the method further includes:
acquiring a test load data set;
dividing a test load data set into a plurality of test subsets according to a plurality of preset time periods;
determining a corresponding target load prediction model according to the time period of the test subset;
inputting the test subsets into corresponding target load prediction models to obtain test results;
calculating the average absolute percentage error and/or the maximum absolute percentage error of the test result of each time period and the actual power load data;
and determining the prediction accuracy of the load prediction model according to the average absolute percentage error and/or the maximum absolute percentage error.
In this embodiment, the computer device tests the target load prediction model through the test load data set, determines the prediction accuracy of the load prediction model based on the test result, and further optimizes the target load prediction model again based on the prediction accuracy, so that the obtained power load prediction result is more accurate.
In a second aspect, there is provided an electrical load prediction apparatus, comprising:
the acquisition module is used for acquiring historical power load data;
the system comprises a dividing module, a storage module and a processing module, wherein the dividing module is used for dividing historical power load data into a plurality of load data sets according to a plurality of preset time periods;
the determining module is used for determining a corresponding target load prediction model according to the time period of each load data set;
and the prediction module is used for inputting each load data set into the corresponding target load prediction model to obtain a power load prediction result of the time period.
In a third aspect, there is provided a computer device comprising a memory storing a computer program and a processor implementing the power load prediction method of any one of the first aspect when the processor executes the computer program.
In a fourth aspect, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the power load prediction method of any one of the first aspects described above.
According to the power load prediction method, the power load prediction device, the computer equipment and the storage medium, historical power load data are obtained by the computer equipment, the historical power load data are divided into a plurality of load data sets according to a plurality of preset time periods, corresponding target load prediction models are determined according to the time periods of the load data sets, and the load data sets are input into the corresponding target load prediction models to obtain power load prediction results of the time periods. In the scheme, because the power load data of different time periods have large differences, in order to carry out power load prediction of each time period in a targeted manner, the power load is predicted by constructing the target load prediction model corresponding to each time period, compared with the traditional algorithm model, the prediction model of each time period can better deal with the periodic variation fluctuation of the power load in different time periods and the influence of multiple factors, has strong practicability, and improves the accuracy of the power load prediction result of each time period.
Drawings
FIG. 1 is a diagram illustrating an internal structure of a computer device according to an embodiment;
FIG. 2 is a flow diagram illustrating a method for predicting a power load according to one embodiment;
FIG. 3 is a flow diagram illustrating a method for predicting a power load according to one embodiment;
FIG. 4 is a flow diagram illustrating a method for predicting a power load according to one embodiment;
FIG. 5 is a flow diagram illustrating a method for predicting a power load according to one embodiment;
FIG. 6 is a flow diagram illustrating a method for predicting a power load according to one embodiment;
FIG. 7 is a flow diagram illustrating a method for predicting a power load according to one embodiment;
FIG. 8 is a flow diagram illustrating a method for predicting a power load according to one embodiment;
FIG. 9 is a block diagram showing the structure of a power load prediction apparatus according to an embodiment;
FIG. 10 is a block diagram showing the structure of a power load prediction apparatus according to an embodiment;
FIG. 11 is a block diagram showing the structure of a power load prediction apparatus according to an embodiment;
fig. 12 is a block diagram showing the structure of a power load prediction apparatus according to an embodiment.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
First, before specifically describing the technical solution of the embodiment of the present disclosure, the technical background on which the embodiment of the present disclosure is based is described. With the rapid development of economy, the demand of society for electric power resources has also been continuously increased. Because the electric energy is difficult to store in a large scale, the power load prediction has important significance for power generation and scheduling planning of power supply enterprises. In recent years, the scales of power grids and power loads are rapidly increasing, and higher requirements are made on the accuracy of Short-term load forecasting (STLF) and the stability of a forecasting model.
The accurate short-term power load prediction can enable power enterprises to master power consumption requirements in time, reasonably arrange production and operation plans of a power generation system, meet social requirements, reduce unnecessary electric energy waste, reduce power generation cost and improve social and economic benefits; meanwhile, due to the relevance of electric power and economic development, important data support can be effectively provided for various policy decisions by timely, quickly and accurately predicting the future power utilization trend. Since the power load is affected by many factors such as economy, weather, time, and area, it is difficult to integrate various factors in prediction, and therefore the accuracy of short-term power load prediction is not easily improved. In a short-term load prediction scene, improving the accuracy of short-term power load prediction and improving the prediction efficiency reaching the short-term power load prediction become problems to be solved urgently at present. In addition, in the present invention, in the context of short-term circuit load prediction, a day is taken as a division target, the day is divided into a plurality of time segments, for example, hours, minutes, and seconds are taken as division bases, training of a weighted combination model of a plurality of prediction models is performed on data of each time segment based on a virtual prediction algorithm to obtain a target short-term load prediction model, and each time segment corresponds to a respective target short-term load prediction model, so that accuracy of a short-term load prediction result of each time segment and a specific technical solution introduced in the following embodiments can be improved.
The power load prediction method provided by the application can be applied to the application environment shown in fig. 1. In one embodiment, a computer device is provided, which may be a server, the internal structure of which may be as shown in fig. 1. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The communication interface of the computer device is used for carrying out wired or wireless communication with an external terminal, and the wireless communication can be realized through WIFI, an operator network, NFC (near field communication) or other technologies. The computer program is executed by a processor to implement a power load prediction method. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.
Those skilled in the art will appreciate that the architecture shown in fig. 1 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
The following describes in detail the technical solutions of the present application and how the technical solutions of the present application solve the above technical problems by embodiments and with reference to the drawings. The following several specific embodiments may be combined with each other, and details of the same or similar concepts or processes may not be repeated in some embodiments. It should be noted that the power load prediction method provided in the embodiments of fig. 2 to 8 of the present application is mainly implemented by a computer device, and may also be a power load prediction apparatus, and the power load prediction apparatus may be a part or all of the computer device by software, hardware, or a combination of software and hardware. In the following method embodiments, the execution subject is a computer device as an example.
In one embodiment, as shown in fig. 2, a power load prediction method is provided, which relates to a process in which a computer device obtains a power load prediction result for a time period by acquiring historical power load data, dividing the historical power load data into a plurality of load data sets according to a plurality of preset time periods, determining a corresponding target load prediction model according to the time period of each load data set, and inputting each load data set into the corresponding target load prediction model, and includes the following steps:
s201, historical power load data is acquired.
The power load data includes power load, weather data, date data, and the like. For example, meteorological data may include temperature, humidity, precipitation, etc.; the date data may include month, week, time of day, etc.
In this embodiment, a computer device obtains historical power load data over a period of time in the past, where the power load may be obtained from a power third party platform; the meteorological data can be acquired from a meteorological third-party platform; date data may be obtained based on a world time database. Alternatively, the computer device may perform data preprocessing on the historical power load data after acquiring the historical power load data, illustratively, performing quantization processing on discrete data, for example, performing quantization processing on the daily data through one-hot coding; or, the data may also be normalized, that is, the data precision and the value range of all the data are processed to the same data precision or value range, which is not limited in this embodiment.
And S202, dividing historical power load data into a plurality of load data sets according to a plurality of preset time periods.
The preset time periods are determined according to the time period corresponding to the target load prediction model. For example, if the target load prediction model includes 8-9 model a, 9-10 model B, and 15-16 model C, the predetermined time period may include 8-9, 9-10, and 15-16.
In the present embodiment, the acquired historical power load data is divided into load data sets according to a preset time period, that is, according to the above example, the load data sets corresponding to the time period from 8 points to 9 points, the load data sets corresponding to the time period from 9 points to 10 points, and the load data sets corresponding to the time period from 15 points to 16 points are obtained through division. Optionally, the historical power load data that does not belong to any one preset time period may be ignored, which is not limited in this embodiment.
And S203, determining a corresponding target load prediction model according to the time period of each load data set.
The target load prediction model is a prediction model corresponding to each time segment. That is, the computer device constructs load prediction models corresponding to respective time periods based on the time periods.
In this embodiment, after obtaining the load data sets of each time period, the computer device determines the target load prediction model corresponding to each load data set according to the time period corresponding to each load prediction model and the time period corresponding to each load data set. Optionally, the computer device may determine, in a traversal matching manner, a target load prediction model matched with a time period corresponding to each load data set, which is not limited in this embodiment.
And S204, inputting each load data set into the corresponding target load prediction model to obtain a power load prediction result of the time period.
In this embodiment, after determining the target load prediction model corresponding to each load data set, the computer device inputs each load data set as a model input into the corresponding target load prediction model to obtain a load prediction result corresponding to each time period. The load prediction result corresponds to the data in the load data set, that is, the load prediction result also includes the prediction result of the power load, the prediction result of the meteorological data, the prediction result of the date data, and the like, which is not limited in this embodiment.
In the power load prediction method, the computer device acquires historical power load data, divides the historical power load data into a plurality of load data sets according to a plurality of preset time periods, determines a corresponding target load prediction model according to the time period of each load data set, and inputs each load data set into the corresponding target load prediction model to obtain the power load prediction result of the time period. In the scheme, because the power load data of different time periods have large differences, in order to carry out power load prediction of each time period in a targeted manner, the power load is predicted by constructing the target load prediction model corresponding to each time period, compared with the traditional algorithm model, the prediction model of each time period can better deal with the periodic variation fluctuation of the power load in different time periods and the influence of multiple factors, has strong practicability, and improves the accuracy of the power load prediction result of each time period.
Optionally, in one embodiment, the target load prediction model includes at least two types of prediction models.
In this embodiment, the target load prediction model of each time segment may be a combined model, and optionally, the target load prediction model may be a combined model weighted by two models, or may be a combined model weighted by three or more models. Illustratively, considering the performance and advantages of the existing prediction model, the weighted combination model of the two models may be a weighted combination model of a Support Vector Machine (SVM) and a Tree-based Pipeline optimization technology (TPOT). The SVM maps training data to a high-dimensional feature space in a non-linear mode, a hyperplane is found in the plane, so that the isolation edge between a positive example and a negative example is maximized, and the generalization capability is strong. TPOT utilizes a genetic algorithm to select features and models, and searches for an optimal data pipeline by exploring thousands of possible pipelines, so that automation of machine learning is realized. The TPOT can automatically complete feature selection, feature preprocessing, feature construction and the like, and can also select a model and optimize parameters. After the search is completed, the content of the specific pipeline when the best performance is obtained is also provided, and the modification and optimization of developers are facilitated.
In this embodiment, the target load prediction model adopts a combined model, and compared with a traditional single model, the combined prediction model can better cope with the periodic variation fluctuation of the power load in different time periods and the influence of multiple factors, has strong practicability, and improves the accuracy of the power load prediction result in each time period.
The computer device needs to construct a target load prediction model corresponding to each time segment. In an embodiment, as shown in fig. 3, the method for constructing the target load prediction model includes:
s301, determining a plurality of preset time periods, and acquiring a sample load data set of each preset time period.
The preset time period can be determined based on actual conditions, for example, through mass data analysis, if the data variation of 8 points-9 points is not large, 8 points-9 points are determined to be a time period; if the data change amount of 8 point-10 point is large and the data change amount is near 9 point, the data change amount can be regarded as a node of data change, and the data change amount can be regarded as two time periods respectively from 8 point to 9 point and from 9 point to 10 point. Alternatively, the time period may be divided in units of hours or minutes, and for the short-term power load, because the load in one day has fluctuation, in this embodiment, the preset time period may be determined in units of 15 minutes. The present embodiment does not limit the division of the time period and the duration of the time period. In this embodiment, the computer device obtains a sample load dataset corresponding to each time period according to the determined multiple preset time periods. Similarly, the sample load data set includes power load, weather data, date data, and the like. For example, meteorological data may include temperature, humidity, precipitation, etc.; the date data may include month, week, time of day, etc.
S302, training the initial load prediction model according to the sample load data set of each preset time period to obtain a target load prediction model corresponding to each preset time period.
Wherein, the initial load prediction models of different time periods have different accuracies because the data fluctuation of the unused time periods is large. In this embodiment, the computer device trains the initial load prediction model according to the sample load data set of each time segment, determines the accuracy of the initial load prediction model of each time segment, and obtains a target load prediction model corresponding to each time segment. Alternatively, in the case where the load prediction model is a weighted combination model of a plurality of different types of models, the weights of the respective models in the load prediction models in different time periods are also different.
In this embodiment, the initial power load prediction models of each time segment are trained in a targeted manner to obtain the target load prediction models of each time segment, and the prediction models of the time segments can better cope with the influence of the periodic variation fluctuation and the multi-factor of the power load in different time segments, so that the accuracy of the power load prediction result of each time segment is improved.
Optionally, when the computer device trains the initial load prediction models in each time period, in an embodiment, as shown in fig. 4, the training the initial load prediction models according to the sample load data sets in each preset time period to obtain the target load prediction models corresponding to each preset time period includes:
s401, for each sample load data set of the preset time period, dividing the sample load data set into first sample load data corresponding to a first sub-time period and second sample load data corresponding to a second sub-time period.
In this embodiment, the fitting accuracy is simply sought as high as possible, and high prediction accuracy is not necessarily guaranteed. Therefore, the present embodiment uses a virtual prediction method to balance the fitting accuracy and the prediction accuracy of the target load prediction model. The virtual prediction refers to the fact that a time period which is later than the historical time period is used as an assumed prediction object, a model and parameters with good effects are selected to be applied to prediction of a future time period, and fitting accuracy and prediction accuracy are well balanced.
The first sample load data corresponding to the first sub-period may be used as virtual prediction data, and the second sample load data corresponding to the second sub-period may be used as an actual result of virtual prediction.
In this embodiment, assuming that the number of time segments per day is T, for the obtained sample load dataset yt, T ∈ [1, T ], a time segment with T ≦ T + n in the future needs to be predicted. Introducing a virtual prediction time period, and dividing the sample load data set into 2 subsets in time order, namely first sample load data of H ═ yt | T ∈ [1, T-m ] } and second sample load data of V ═ yt | T ∈ [ T-m-1, T ] }.
S402, inputting the first sample load data into the initial load prediction model to obtain a prediction result.
In the embodiment, historical simulation is firstly carried out, data in the set H is analyzed and modeled, an initial load prediction model A reaching certain fitting precision is selected, and virtual prediction is carried out on a time period T-m-1 which is less than or equal to T and less than or equal to T by using the initial load prediction model A, so that a prediction result is obtained.
And S403, adjusting parameters of the initial load prediction model according to the prediction result and the second sample load data to obtain a target load prediction model.
In the embodiment, according to the prediction result and the set V, that is, the second sample load data, the parameters of the initial load prediction model a are adjusted to make the result of the virtual prediction as close as possible to the data in the set V, and finally, the adjusted initial load prediction model a and the optimized parameters thereof are used to make a real prediction on the future time period T ≦ T + n.
In the embodiment, the data of the prediction time period in the virtual prediction is known, and the accuracy of the prediction model can be checked, so that the accuracy of the real prediction is improved.
In the case where the load prediction model is a weighted combination model of a plurality of models, the computer device needs to determine the weight values of the respective models in the initial load prediction model for the respective time periods. In one embodiment, as shown in fig. 5, the initial load prediction model includes at least two types of prediction models, and the method further includes:
and S501, calculating the residual square sum of the prediction result and the second sample load data according to the prediction result output by the initial load prediction model of each preset time period.
In the present embodiment, by way of example, it is assumed that the number of time segments per day is T, and the actual load of the nth day and the tth time segment is historically Ln,t(n ═ C-1, C-2.; (T ═ 1, 2.., T), C is the current day. Assuming that M models are used for load prediction, the weight of the ith model in the t time period is wi,t(ii) a Predicted result of the ith is Li′,C+F,tAnd F is the number of days to be predicted.
Performing virtual test on a virtual test sample set of historical recent N days, wherein the virtual prediction result of the ith model on the nth day is L'i,n,t(i＝1,2,...,M；t＝1,2,...,T；n＝C-1,C-1,...,n＝C-N)。
Wherein the content of the first and second substances,
in the embodiment, the prediction result of virtual prediction is carried out by using the initial load prediction model in the t time period, and the residual square sum Z of the prediction result and the second sample load data is determinedtThe calculation formula is as follows:
s502, determining the weight of the initial load prediction model after parameters are adjusted according to the residual sum of squares and a preset optimization objective function to obtain a target load prediction model.
In this embodiment, the preset optimization objective function may be an optimization function with the minimum sum of squared residuals, that is, the optimization objective function may be expressed as:
wi,t≥0,i＝1,2,...,M
in this embodiment, the optimization problem is solved by a nonlinear specification method by substituting the sum of squared residuals into the above-mentioned optimization objective function, and finally, the optimal weight of each model in the initial load prediction model after the parameters are adjusted can be determined, so as to obtain the target load prediction model.
In this embodiment, the computer device may optimize and solve to obtain the optimal weight of each model in the initial load prediction model by setting an optimized objective function and calculating the sum of squares of residuals of the prediction result and the second sample load data, so as to obtain the target load prediction model, and an output result of the optimized and solved target load prediction model is more accurate.
After the computer device acquires the original load data set, no matter training of the load prediction model or prediction of the power load is performed, data preprocessing needs to be performed on the original load data set. In an embodiment, as shown in fig. 6, the determining a plurality of preset time periods and obtaining a sample load data set for each preset time period includes:
s601, acquiring an original load data set of each preset time period.
In the present embodiment, similarly, the raw load data set acquired by the computer device includes the power load, the weather data, and the date data and the like. For example, meteorological data may include temperature, humidity, precipitation, etc.; the date data may include month, week, time of day, etc.
The change of the power load mainly depends on the regularity of production and life of people, the power load of a working day is obviously improved due to production economic activities in relative rest days, the power load of different months such as summer and winter also brings about the change of the power consumption due to temperature regulation requirements, so that the short-term power load has obvious timeliness and periodicity, the working day and the rest day have respective similarity, the working day and the rest day have corresponding similarity in different weeks and the same week type day, and the monthly degree and the season also have similarity, so that the time characteristics, the week type day, the month and other factors need to be taken as characteristic data for key consideration.
Meteorological factors have obvious influence on short-term power load, and when the weather is changed violently, a large amount of heating and cooling loads are generated; when rainfall changes, the rainfall has important influence on agricultural machinery irrigation load, so that meteorological data such as temperature, humidity, rainfall and other indexes become important characteristic data in short-term power load prediction.
And S602, carrying out data preprocessing on the original load data set to obtain a sample load data set of each preset time period.
The data preprocessing comprises at least one of data feature extraction, data missing value processing, data normalization processing, numeralization processing and data feature and power load association degree analysis.
In this embodiment, the computer device needs to perform data preprocessing on the raw load data set to normalize the input data of the load prediction model.
Data feature extraction: the process of data feature extraction includes determining factors related to the electrical load. Because the power load is subjected to the combined action of various factors, the effect of the multiple-day accumulation effect and the combined effect of the multiple factors need to be considered in addition to the action of the single factor. Such as continuous rainfall and sudden rainfall on the same day, have a significant difference in the effect on the daily load, so that the cumulative factors adjacent to the predicted time point are also required as input characteristic data. For this reason, for predicting the load demand after one day, the corresponding characteristic data is selected in the present embodiment, as shown in table 1.
TABLE 1 short term Power load characterization data
Data missing value processing: because the acquisition periods for different influencing factors are different, the data of different dimensions have different time granularities, for example, the power load data is acquired every 15 minutes, 96 pieces of acquired data can exist every day, and the meteorological data, including temperature, humidity and precipitation, are recorded every hour, and 24 pieces of acquired data every day. Therefore, all indexes need to be unified to the same time scale, the indexes with large time granularity are automatically supplemented to the indexes with small time granularity, and the data processing process is data missing value processing. Exemplarily, the computer device needs to align the power load data and the meteorological data according to time, and because the meteorological data are acquired by hours and the time granularity is large, the meteorological data of every 15 minutes are automatically filled by taking the value of the nearest time distance point, that is, 4 pieces of meteorological data in each hour in the obtained meteorological data are the same.
Data normalization processing: because the dimensions of each dimension data are different, the values of different dimensions need to be unified to a specific interval, so that each index can have numerical comparability, and the data processing process is data normalization processing. Optionally, in this embodiment, the computer device performs data normalization using a min-max normalization method. The formula is as follows:
where x is the input data, xmax、xminThe maximum value and the minimum value of the input data are respectively, and x' is the data after normalization processing.
And (3) numerical processing: due to the discreteness of the date type data, one-hot encoding is adopted for preprocessing, 7-dimensional, 24-dimensional and 12-dimensional one-hot encoding is respectively adopted for the sunday type, the time characteristic and the month, and for the sunday type data, for example, initial 7-dimensional data can be set as [0000000], data corresponding to monday as [1000000], data corresponding to tuesday as [0100000] and data corresponding to wednesday as [0010000 ]; for month type data, initial 12-dimensional data may be set as [00000000000000], data corresponding to january as [10000000000000], data corresponding to february as [01000000000000], and data corresponding to march as [00100000000000], in this embodiment, the digitization processing of the corresponding dimensional data of the discrete data is performed from left to right, of course, the digitization processing of the corresponding dimensional data of the discrete data may be performed from right to left, as long as the digitization calibration can be performed on the discrete data, which specification embodiment is not limited thereto.
Analyzing the relevance of the data characteristics and the power load: the influence factors of the power load are more, so that the characteristics with larger influence need to be selected according to the data condition for training and prediction, thereby reducing redundant data characteristics, reducing the complexity of a model and improving the prediction precision. The invention analyzes the association between each factor and the power load by adopting distance analysis. The distance analysis measures the difference between the variables, calculates the generalized distance between different variables, and adopts the Euclidean distance, and the calculation formula of the Euclidean distance is as follows:
wherein (x)i，yi) The coordinates of the ith data feature in a Euclidean coordinate system; n is the number of data features.
In this embodiment, the euclidean distance is used for the difference degree analysis of each data feature, and the data features with shorter distance and greater relevance are selected according to the calculation result for the subsequent model training and prediction.
In this embodiment, data feature extraction, data missing value processing, data normalization processing, digitization processing, and data feature and power load association degree analysis are performed on the original load data set, so as to obtain input data of the load prediction model under the normalization processing, and thus the obtained output result of the load prediction model is more accurate.
After obtaining the load prediction models of the respective time periods, optionally, the computer device may further perform model testing on the load prediction models of the respective time periods. In one embodiment, as shown in fig. 7, the method further includes:
and S701, acquiring a test load data set.
In this embodiment, similar to the step 201 described above, a test load dataset is obtained in this embodiment, which also includes power loads, meteorological data, and date data, etc. The computer device may obtain the test load dataset from a preset storage space, or may obtain the open-source test load dataset from a third-party platform.
And S702, dividing the test load data set into a plurality of test subsets according to a plurality of preset time periods.
In this embodiment, similar to the step 202, the test load data set is divided into a plurality of test subsets according to a plurality of preset time periods, which is not described in detail in this embodiment.
And S703, determining a corresponding target load prediction model according to the time period of the test subset.
In this embodiment, similar to step 203, the target load prediction model corresponding to each test subset is determined according to the time period of the test subset and the time period of each target load prediction model, which is not described in detail in this embodiment.
And S704, inputting the test subset into the corresponding target load prediction model to obtain a test result.
In this embodiment, similar to the step 204, each test subset is input to the corresponding target load prediction model to obtain a test result, which is not described in detail in this embodiment.
S705, calculating the average absolute percentage error and/or the maximum absolute percentage error of the test result of each time period and the actual power load data.
In this embodiment, the average absolute percentage error (MAPE) and/or the maximum absolute percentage error (max absolute percentage error, MaxE) are/is used as a test index to calculate the prediction accuracy of the target load prediction model.
Alternatively, the MAPE and MaxE are calculated as follows:
wherein x isiIs the actual power load value, xiIs a negative output by the target load prediction modelAnd (4) a load test value.
And S706, determining the prediction accuracy of the load prediction model according to the average absolute percentage error and/or the maximum absolute percentage error.
In this embodiment, the prediction accuracy of the load prediction model is determined according to the values of MAPE and MaxE, and a smaller value of MAPE and MaxE indicates a more accurate load prediction.
In this embodiment, the computer device tests the target load prediction model through the test load data set, determines the prediction accuracy of the load prediction model based on the test result, and further optimizes the target load prediction model again based on the prediction accuracy, so that the obtained power load prediction result is more accurate.
To better explain the above method, as shown in fig. 8, the present embodiment provides a power load prediction method, which specifically includes:
s101, acquiring an original load data set of each preset time period;
s102, carrying out data preprocessing on the original load data set to obtain a sample load data set of each preset time period;
s103, dividing the sample load data set into first sample load data corresponding to a first sub-time period and second sample load data corresponding to a second sub-time period aiming at the sample load data set of each preset time period;
s104, inputting the first sample load data into an initial load prediction model to obtain a prediction result;
s105, adjusting parameters of the initial load prediction model according to the prediction result and the second sample load data to obtain a target load prediction model;
s106, calculating the residual square sum of the prediction result and the second sample load data according to the prediction result output by the initial load prediction model of each preset time period;
s107, determining the weight of the initial load prediction model after parameters are adjusted according to the residual sum of squares and a preset optimization objective function to obtain a target load prediction model;
s108, acquiring a test load data set;
s109, dividing the test load data set into a plurality of test subsets according to a plurality of preset time periods;
s110, determining a corresponding target load prediction model according to the time period of the test subset;
s111, inputting the test subsets into corresponding target load prediction models to obtain test results;
s112, calculating the average absolute percentage error and/or the maximum absolute percentage error between the test result of each time period and the actual power load data;
s113, determining the prediction accuracy of the load prediction model according to the average absolute percentage error and/or the maximum absolute percentage error;
s114, acquiring historical power load data;
s115, dividing historical power load data into a plurality of load data sets according to a plurality of preset time periods;
s116, determining a corresponding target load prediction model according to the time period of each load data set;
and S117, inputting each load data set into the corresponding target load prediction model to obtain a power load prediction result of the time slot.
In the embodiment, since the power load data of different time periods have large differences, in order to predict the power load of each time period in a targeted manner, the power load is predicted by constructing the target load prediction model corresponding to each time period, and in the embodiment, when the target load prediction model corresponding to each time period is constructed, a combination model of multiple types of prediction models is adopted, compared with a traditional algorithm and a single model, the combination prediction model of the time periods can better cope with the influence of periodic variation fluctuation and multiple factors of the power load in different time periods, has strong practicability, and improves the accuracy of the power load prediction result of each time period.
The implementation principle and technical effect of the power load prediction method provided by the above embodiment are similar to those of the above embodiment, and are not described herein again.
It should be understood that although the various steps in the flow charts of fig. 2-8 are shown in order as indicated by the arrows, the steps are not necessarily performed in order as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least some of the steps in fig. 2-8 may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, which are not necessarily performed in sequence, but may be performed in turn or alternately with other steps or at least some of the other steps.
In one embodiment, as shown in fig. 9, there is provided an electric load prediction apparatus including: the prediction module comprises an acquisition module 01, a division module 02, a determination module 03 and a prediction module 04, wherein:
the acquisition module 01 is used for acquiring historical power load data;
the dividing module 02 is used for dividing historical power load data into a plurality of load data sets according to a plurality of preset time periods;
the determining module 03 is configured to determine a corresponding target load prediction model according to a time period of each load data set;
and the prediction module 04 is used for inputting each load data set into the corresponding target load prediction model to obtain a power load prediction result of the time period.
In one embodiment, the target load prediction model includes at least two types of prediction models.
In one embodiment, as shown in fig. 10, the power load prediction apparatus further includes a building module 05;
the building module 05 is used for determining a plurality of preset time periods and acquiring a sample load data set of each preset time period; and training the initial load prediction model according to the sample load data set of each preset time period to obtain a target load prediction model corresponding to each preset time period.
In one embodiment, the root building module 05 is configured to, for each sample load data set of a preset time period, divide the sample load data set into first sample load data corresponding to a first sub-time period and second sample load data corresponding to a second sub-time period; inputting the first sample load data into an initial load prediction model to obtain a prediction result; and adjusting parameters of the initial load prediction model according to the prediction result and the second sample load data to obtain a target load prediction model.
In one embodiment, the initial load prediction model includes at least two types of prediction models, as shown in fig. 11, the power load prediction apparatus further includes an adjusting module 06;
the adjusting module 06 is configured to calculate, for the prediction result output by the initial load prediction model in each preset time period, a residual square sum of the prediction result and the second sample load data; and determining the weight of the initial load prediction model after the parameters are adjusted according to the residual sum of squares and a preset optimization objective function to obtain a target load prediction model.
In one embodiment, the dividing module 02 is configured to obtain an original load data set of each preset time period; carrying out data preprocessing on the original load data set to obtain a sample load data set of each preset time period; the data preprocessing comprises at least one of data feature extraction, data missing value processing, data normalization processing, numeralization processing and data feature and power load association degree analysis.
In one embodiment, as shown in fig. 12, the power load prediction apparatus further includes a test module 07;
the test module 07 is used for acquiring a test load data set; dividing a test load data set into a plurality of test subsets according to a plurality of preset time periods; determining a corresponding target load prediction model according to the time period of the test subset; inputting the test subsets into corresponding target load prediction models to obtain test results; calculating the average absolute percentage error and/or the maximum absolute percentage error of the test result of each time period and the actual power load data; and determining the prediction accuracy of the load prediction model according to the average absolute percentage error and/or the maximum absolute percentage error.
For specific limitations of the power load prediction apparatus, reference may be made to the above limitations of the power load prediction method, which are not described herein again. Each module in the above power load prediction apparatus may be wholly or partially implemented by software, hardware, or a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program:
acquiring historical power load data;
dividing historical power load data into a plurality of load data sets according to a plurality of preset time periods;
determining a corresponding target load prediction model according to the time period of each load data set;
and inputting each load data set into a corresponding target load prediction model to obtain a power load prediction result of a time period.
The implementation principle and technical effect of the computer device provided by the above embodiment are similar to those of the above method embodiment, and are not described herein again.
In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of:
acquiring historical power load data;
dividing historical power load data into a plurality of load data sets according to a plurality of preset time periods;
determining a corresponding target load prediction model according to the time period of each load data set;
and inputting each load data set into a corresponding target load prediction model to obtain a power load prediction result of a time period.
The implementation principle and technical effect of the computer-readable storage medium provided by the above embodiments are similar to those of the above method embodiments, and are not described herein again.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method of predicting a power load, the method comprising:
acquiring historical power load data;
dividing the historical power load data into a plurality of load data sets according to a plurality of preset time periods;
determining a corresponding target load prediction model according to the time period of each load data set;
and inputting each load data set into a corresponding target load prediction model to obtain a power load prediction result of the time period.
2. The method of claim 1, wherein the target load prediction model comprises at least two types of prediction models.
3. The method of claim 1, wherein the target load prediction model is constructed by a method comprising:
determining a plurality of preset time periods, and acquiring a sample load data set of each preset time period;
and training an initial load prediction model according to the sample load data set of each preset time period to obtain a target load prediction model corresponding to each preset time period.
4. The method of claim 3, wherein the training an initial load prediction model according to the sample load data set of each of the preset time periods to obtain a target load prediction model corresponding to each of the preset time periods comprises:
for each sample load data set of a preset time period, dividing the sample load data set into first sample load data corresponding to a first sub-time period and second sample load data corresponding to a second sub-time period;
inputting the first sample load data into the initial load prediction model to obtain a prediction result;
and adjusting parameters of the initial load prediction model according to the prediction result and the second sample load data to obtain the target load prediction model.
5. The method of claim 4, wherein the initial load prediction model comprises at least two types of prediction models, the method further comprising:
calculating the square sum of residual errors of the prediction result and the second sample load data according to the prediction result output by the initial load prediction model of each preset time period;
and determining the weight of the initial load prediction model after parameters are adjusted according to the residual sum of squares and a preset optimization objective function to obtain the target load prediction model.
6. The method of claim 3, wherein determining a plurality of predetermined time periods, obtaining a sample load dataset for each of the predetermined time periods, comprises:
acquiring an original load data set of each preset time period;
performing data preprocessing on the original load data set to obtain a sample load data set of each preset time period;
the data preprocessing comprises at least one of data feature extraction, data missing value processing, data normalization processing, numeralization processing and data feature and power load association degree analysis.
7. The method according to any one of claims 1-6, further comprising:
acquiring a test load data set;
dividing the test load data set into a plurality of test subsets according to a plurality of preset time periods;
determining a corresponding target load prediction model according to the time period of the test subset;
inputting the test subset into a corresponding target load prediction model to obtain a test result;
calculating the average absolute percentage error and/or the maximum absolute percentage error of the test result of each time period and the actual power load data;
and determining the prediction accuracy of the load prediction model according to the average absolute percentage error and/or the maximum absolute percentage error.
8. An electrical load prediction apparatus, the apparatus comprising:
the acquisition module is used for acquiring historical power load data;
the dividing module is used for dividing the historical power load data into a plurality of load data sets according to a plurality of preset time periods;
the determining module is used for determining a corresponding target load prediction model according to the time period of each load data set;
and the prediction module is used for inputting each load data set into a corresponding target load prediction model to obtain a power load prediction result of the time period.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110503715.9A CN113255973A (en) | 2021-05-10 | 2021-05-10 | Power load prediction method, power load prediction device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110503715.9A CN113255973A (en) | 2021-05-10 | 2021-05-10 | Power load prediction method, power load prediction device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113255973A true CN113255973A (en) | 2021-08-13 |
Family
ID=77222354
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110503715.9A Pending CN113255973A (en) | 2021-05-10 | 2021-05-10 | Power load prediction method, power load prediction device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113255973A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113837503A (en) * | 2021-11-29 | 2021-12-24 | 新风光电子科技股份有限公司 | Power grid load prediction method and system based on artificial intelligence |
| CN115660233A (en) * | 2022-12-26 | 2023-01-31 | 尚特杰电力科技有限公司 | Photovoltaic power prediction method and device, electronic equipment and storage medium |
| CN116979531A (en) * | 2023-09-25 | 2023-10-31 | 山西京能售电有限责任公司 | Novel energy data monitoring method and method for monitoring auxiliary power market |
| CN117154740A (en) * | 2023-10-31 | 2023-12-01 | 国网浙江省电力有限公司宁波供电公司 | Load regulation and control method and device for heat accumulating type electric heating participating power distribution network |
| CN117236531A (en) * | 2023-11-16 | 2023-12-15 | 南京大全电气研究院有限公司 | Load prediction method and device, electronic equipment and storage medium |
| CN117318055A (en) * | 2023-12-01 | 2023-12-29 | 山东理工昊明新能源有限公司 | Power load prediction model processing method and device, electronic equipment and storage medium |
| CN117494797A (en) * | 2023-11-21 | 2024-02-02 | 广东电网有限责任公司佛山供电局 | Photovoltaic energy storage method and system |
| CN118070560A (en) * | 2024-04-17 | 2024-05-24 | 华电电力科学研究院有限公司 | Fan tower barrel-shaped state early warning model construction method, early warning method and system |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004094437A (en) * | 2002-08-30 | 2004-03-25 | Fuji Electric Holdings Co Ltd | Data prediction method and data prediction system |
| JP2014157457A (en) * | 2013-02-15 | 2014-08-28 | Nec Corp | Prediction device and prediction method |
| CN105989420A (en) * | 2015-02-12 | 2016-10-05 | 西门子公司 | Method of determining user electricity consumption behavior features, method of predicting user electricity consumption load and device |
| CN106779227A (en) * | 2016-12-23 | 2017-05-31 | 国网浙江省电力公司 | A kind of load self-adapting Forecasting Methodology and device |
| CN108022004A (en) * | 2017-11-16 | 2018-05-11 | 广东电网有限责任公司信息中心 | A kind of adaptive weighting training method of multi-model weighted array Forecasting Power System Load |
| CN109376972A (en) * | 2019-01-01 | 2019-02-22 | 中南大学 | A kind of wisdom Power Network Short-Term Electric Load Forecasting method based on block cluster |
| CN109376971A (en) * | 2018-12-29 | 2019-02-22 | 北京中电普华信息技术有限公司 | A kind of load curve forecasting method and system towards power consumer |
| CN109508835A (en) * | 2019-01-01 | 2019-03-22 | 中南大学 | A kind of wisdom Power Network Short-Term Electric Load Forecasting method of integrated environment feedback |
| CN110245801A (en) * | 2019-06-19 | 2019-09-17 | 中国电力科学研究院有限公司 | A kind of Methods of electric load forecasting and system based on combination mining model |
| CN111210056A (en) * | 2019-12-25 | 2020-05-29 | 深圳供电局有限公司 | Electricity price scheme determination method and device, computer equipment and storage medium |
| CN111709548A (en) * | 2020-04-27 | 2020-09-25 | 深圳华工能源技术有限公司 | Power consumer load prediction method, device, equipment and storage medium based on support vector machine |
| CN112215398A (en) * | 2020-04-27 | 2021-01-12 | 深圳华工能源技术有限公司 | Power consumer load prediction model establishing method, device, equipment and storage medium |
| CN112215442A (en) * | 2020-11-27 | 2021-01-12 | 中国电力科学研究院有限公司 | Method, system, device and medium for predicting short-term load of power system |
| CN112329997A (en) * | 2020-10-26 | 2021-02-05 | 国网河北省电力有限公司雄安新区供电公司 | Power demand load prediction method and system, electronic device, and storage medium |
-
2021
- 2021-05-10 CN CN202110503715.9A patent/CN113255973A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2004094437A (en) * | 2002-08-30 | 2004-03-25 | Fuji Electric Holdings Co Ltd | Data prediction method and data prediction system |
| JP2014157457A (en) * | 2013-02-15 | 2014-08-28 | Nec Corp | Prediction device and prediction method |
| CN105989420A (en) * | 2015-02-12 | 2016-10-05 | 西门子公司 | Method of determining user electricity consumption behavior features, method of predicting user electricity consumption load and device |
| CN106779227A (en) * | 2016-12-23 | 2017-05-31 | 国网浙江省电力公司 | A kind of load self-adapting Forecasting Methodology and device |
| CN108022004A (en) * | 2017-11-16 | 2018-05-11 | 广东电网有限责任公司信息中心 | A kind of adaptive weighting training method of multi-model weighted array Forecasting Power System Load |
| CN109376971A (en) * | 2018-12-29 | 2019-02-22 | 北京中电普华信息技术有限公司 | A kind of load curve forecasting method and system towards power consumer |
| CN109376972A (en) * | 2019-01-01 | 2019-02-22 | 中南大学 | A kind of wisdom Power Network Short-Term Electric Load Forecasting method based on block cluster |
| CN109508835A (en) * | 2019-01-01 | 2019-03-22 | 中南大学 | A kind of wisdom Power Network Short-Term Electric Load Forecasting method of integrated environment feedback |
| CN110245801A (en) * | 2019-06-19 | 2019-09-17 | 中国电力科学研究院有限公司 | A kind of Methods of electric load forecasting and system based on combination mining model |
| CN111210056A (en) * | 2019-12-25 | 2020-05-29 | 深圳供电局有限公司 | Electricity price scheme determination method and device, computer equipment and storage medium |
| CN111709548A (en) * | 2020-04-27 | 2020-09-25 | 深圳华工能源技术有限公司 | Power consumer load prediction method, device, equipment and storage medium based on support vector machine |
| CN112215398A (en) * | 2020-04-27 | 2021-01-12 | 深圳华工能源技术有限公司 | Power consumer load prediction model establishing method, device, equipment and storage medium |
| CN112329997A (en) * | 2020-10-26 | 2021-02-05 | 国网河北省电力有限公司雄安新区供电公司 | Power demand load prediction method and system, electronic device, and storage medium |
| CN112215442A (en) * | 2020-11-27 | 2021-01-12 | 中国电力科学研究院有限公司 | Method, system, device and medium for predicting short-term load of power system |
Non-Patent Citations (4)
| Title |
|---|
| 彭安华: "《面向机械制造过程的模糊多准则决策方法研究》", 31 January 2018, 合肥工业大学出版社, pages: 14 * |
| 蔡小华;吕干云;: "日电力负荷的分时段多模型组合预测", 计算机系统应用, no. 03, pages 98 - 101 * |
| 谢琼娟: "《自动机器学习入门与实践：使用Python》", 31 December 2019, 华中科技大学出版社, pages: 21 - 22 * |
| 高嘉豪;全心雨;陈超?;余少华;: "基于用户分类的配电网负荷预测方法", 科学技术创新, no. 28 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113837503A (en) * | 2021-11-29 | 2021-12-24 | 新风光电子科技股份有限公司 | Power grid load prediction method and system based on artificial intelligence |
| CN115660233A (en) * | 2022-12-26 | 2023-01-31 | 尚特杰电力科技有限公司 | Photovoltaic power prediction method and device, electronic equipment and storage medium |
| CN116979531A (en) * | 2023-09-25 | 2023-10-31 | 山西京能售电有限责任公司 | Novel energy data monitoring method and method for monitoring auxiliary power market |
| CN116979531B (en) * | 2023-09-25 | 2023-12-12 | 山西京能售电有限责任公司 | Novel energy data monitoring method and method for monitoring auxiliary power market |
| CN117154740A (en) * | 2023-10-31 | 2023-12-01 | 国网浙江省电力有限公司宁波供电公司 | Load regulation and control method and device for heat accumulating type electric heating participating power distribution network |
| CN117154740B (en) * | 2023-10-31 | 2024-01-26 | 国网浙江省电力有限公司宁波供电公司 | Load regulation and control method and device for heat accumulating type electric heating participating power distribution network |
| CN117236531A (en) * | 2023-11-16 | 2023-12-15 | 南京大全电气研究院有限公司 | Load prediction method and device, electronic equipment and storage medium |
| CN117494797A (en) * | 2023-11-21 | 2024-02-02 | 广东电网有限责任公司佛山供电局 | Photovoltaic energy storage method and system |
| CN117318055A (en) * | 2023-12-01 | 2023-12-29 | 山东理工昊明新能源有限公司 | Power load prediction model processing method and device, electronic equipment and storage medium |
| CN117318055B (en) * | 2023-12-01 | 2024-03-01 | 山东理工昊明新能源有限公司 | Power load prediction model processing method and device, electronic equipment and storage medium |
| CN118070560A (en) * | 2024-04-17 | 2024-05-24 | 华电电力科学研究院有限公司 | Fan tower barrel-shaped state early warning model construction method, early warning method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113255973A (en) | Power load prediction method, power load prediction device, computer equipment and storage medium | |
| CN111091196B (en) | Passenger flow data determination method and device, computer equipment and storage medium | |
| CN103218675A (en) | Short-term load prediction method based on clustering and sliding window | |
| CN112734135B (en) | Power load prediction method, intelligent terminal and computer readable storage medium | |
| CN112330077B (en) | Power load prediction method, power load prediction device, computer equipment and storage medium | |
| Dong et al. | Applying the ensemble artificial neural network-based hybrid data-driven model to daily total load forecasting | |
| CN111680841B (en) | Short-term load prediction method, system and terminal equipment based on principal component analysis | |
| CN114065653A (en) | Construction method of power load prediction model and power load prediction method | |
| CN114330935A (en) | New energy power prediction method and system based on multiple combined strategy integrated learning | |
| CN113743673A (en) | Power load prediction method during typhoon | |
| CN111815039A (en) | Weekly scale wind power probability prediction method and system based on weather classification | |
| CN112036598A (en) | Charging pile use information prediction method based on multi-information coupling | |
| CN109858668B (en) | Coordination prediction method for power load region in thunder and lightning climate | |
| CN113887809A (en) | Power distribution network supply and demand balance method, system, medium and computing equipment under double-carbon target | |
| CN114186733A (en) | Short-term load prediction method and device | |
| Khan et al. | Comparison of electric vehicle load forecasting across different spatial levels with incorporated uncertainty estimation | |
| Alharbi et al. | Short-term wind speed and temperature forecasting model based on gated recurrent unit neural networks | |
| CN116316617B (en) | Multi-station intelligent fusion new energy generation power region prediction method and system | |
| CN117791618A (en) | Power scheduling method, device, computer equipment and storage medium | |
| Zhang et al. | Building energy consumption prediction based on temporal-aware attention and energy consumption states | |
| CN107464015B (en) | Short-term load with fusing mechanism and electric quantity prediction method based on electricity utilization behavior | |
| Kulikova et al. | Optimization algorithms to solve problems of comprehensive electrification based on renewable energy sources | |
| CN115549082A (en) | Power load prediction method based on load mining and LSTM neural network | |
| CN115358476A (en) | Demand response potential prediction method, demand response potential prediction device, computer equipment and storage medium | |
| CN114819337A (en) | Multi-task learning-based comprehensive energy system multi-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 |