CN112257928A - Short-term power load probability prediction method based on CNN and quantile regression - Google Patents
Short-term power load probability prediction method based on CNN and quantile regression Download PDFInfo
- Publication number
- CN112257928A CN112257928A CN202011140762.3A CN202011140762A CN112257928A CN 112257928 A CN112257928 A CN 112257928A CN 202011140762 A CN202011140762 A CN 202011140762A CN 112257928 A CN112257928 A CN 112257928A
- Authority
- CN
- China
- Prior art keywords
- data
- prediction
- short
- quantile
- probability density
- 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 65
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 50
- 238000012549 training Methods 0.000 claims abstract description 25
- 238000012360 testing method Methods 0.000 claims abstract description 13
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 13
- 238000012545 processing Methods 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 238000004140 cleaning Methods 0.000 claims description 3
- 230000001502 supplementing effect Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 8
- 238000013528 artificial neural network Methods 0.000 description 4
- 238000011156 evaluation Methods 0.000 description 4
- 238000009826 distribution Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 239000004579 marble Substances 0.000 description 3
- 230000007547 defect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 108090000672 Annexin A5 Proteins 0.000 description 1
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000556 factor analysis Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000009776 industrial production Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008689 nuclear function Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000000611 regression analysis Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/045—Combinations of networks
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Data Mining & Analysis (AREA)
- Economics (AREA)
- Strategic Management (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Health & Medical Sciences (AREA)
- Marketing (AREA)
- Software Systems (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Operations Research (AREA)
- Evolutionary Computation (AREA)
- Pure & Applied Mathematics (AREA)
- Molecular Biology (AREA)
- Databases & Information Systems (AREA)
- Algebra (AREA)
- Probability & Statistics with Applications (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Primary Health Care (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
Abstract
The invention discloses a short-term power load probability prediction method based on CNN and quantile regression, which comprises the following steps: the method comprises the following steps: collecting power load data, and determining key influence factors according to the relevance of the power load data and external influence factors; step two: preprocessing data, and dividing input data into training set data and test set data; step three: performing short-term load probability density prediction model training based on a convolutional neural network and quantile regression by using the training set data in the step two to obtain a trained short-term load probability density prediction model based on the convolutional neural network and quantile regression; step four: inputting the test data into a trained QRCNN model to obtain predicted values under different quantile points; step five: and taking the predicted values under different quantiles as input, and performing load probability density prediction by using a kernel density estimation method under different confidence coefficients to obtain a prediction interval and a probability density curve.
Description
Technical Field
The invention relates to the technical field of short-term power load prediction, in particular to a short-term load probability prediction method based on a convolutional neural network and quantile regression.
Background
The electric energy is an indispensable part in daily life and industrial production, because the electric energy has the properties of real-time property and difficulty in mass storage, reasonable prediction of the electric load becomes a necessary premise for maintaining stable operation of the electric power system, and meanwhile, the problem of accurately predicting the electric power system is also the problem of the electric power system.
According to previous studies, deterministic load prediction methods can be roughly divided into two categories: statistical models and machine learning models. The statistical models mainly include an autoregressive moving average (ARMA) model, an Exponential Smoothing (ES) model, a Multiple Linear Regression (MLR) model and a semi-parametric additive model. The machine learning model mainly comprises Support Vector Regression (SVR), Random Forest (RF) and Artificial Neural Network (ANN). However, since the conventional point prediction method provides a single prediction estimation value, and the high uncertainty of the power load cannot be accurately measured, more and more scholars convert the research field from deterministic prediction to interval prediction. Many scholars have begun to combine quantile regression methods with machine learning models to obtain high quality interval load predictions. Then, more comprehensive prediction information can be obtained by further estimating the probability density, and methods for estimating the probability density of a given data set mainly fall into two categories: parametric estimation methods and non-parametric estimation methods. Probability density prediction is a prediction method which can provide the most load prediction information at present and has the most comprehensive reference value.
Therefore, a short-term load probability prediction method based on a convolutional neural network and quantile regression is expected to solve the problems in the prior art.
Disclosure of Invention
The invention discloses a short-term load probability prediction method based on a convolutional neural network and quantile regression, which comprises the following steps of:
the method comprises the following steps: collecting power load data, and determining input data according to the relevance of the power load data and external influence factors;
step two: preprocessing input data, and segmenting the input data into a training set and a test set;
step three: performing short-term load probability density prediction model training based on a convolutional neural network and quantile regression by using the training set data in the step two to obtain a trained short-term load probability density prediction model based on the convolutional neural network and quantile regression;
step four: inputting the test set data in the step two into a trained short-term load probability density prediction model based on a convolutional neural network and quantile regression to obtain predicted values under different quantile points;
step five: and taking the predicted values under different quantiles as input, and performing load probability density prediction by using a kernel density estimation method under different confidence coefficients to obtain a prediction interval and a probability density curve.
Preferably, the step one input data includes load value and temperature.
Preferably, the input data preprocessing of the second step adopts a MinMax method to supplement missing data values and clean abnormal values, and then normalization processing is carried out on the data by using a formula (1) so that a data value domain is converted into [0,1]]Wherein X isnormIs normalized data, x is raw data, xmin、xmaxMaximum and minimum values for data:
preferably, the first and second electrodes are formed of a metal,
the short-term load probability density prediction model in the third step is obtained by training in the following mode:
and taking the quantile loss as a loss function of the convolutional neural network, and optimizing the parameters of the convolutional neural network to obtain the short-term load probability density prediction model.
Preferably, the quantile loss is output once through a model for the prediction results under different quantile points, the quantile loss is shown in formula (2),as a function of the quantile loss function,denotes the j (th)thThe predicted value of the sample point under the i quantile, and q represents the number of quantiles: :
preferably, the kernel density estimation method in the fifth step adopts an Epanechnikov kernel as a kernel function, and intercepts probability density data of a corresponding prediction interval according to the requirements of different confidence degrees.
Preferably, a short-term power load probability prediction apparatus based on CNN and quantile regression includes:
the acquisition module is used for acquiring power load data and determining input data according to the correlation between the power load data and external influence factors;
the processing module is used for preprocessing input data and dividing the input data into training set data and test set data;
the training module is used for carrying out short-term load probability density prediction model training based on a convolutional neural network and quantile regression by using the training set data to obtain a trained short-term load probability density prediction model based on the convolutional neural network and the quantile regression;
the input module is used for inputting the test set data into a trained short-term load probability density prediction model based on a convolutional neural network and quantile regression to obtain predicted values under different quantile points;
and the prediction module is used for taking the predicted values under different quantiles as input and performing load probability density prediction by using a kernel density estimation method under different confidence coefficients to obtain a prediction interval and a probability density curve.
Preferably, the prediction module is specifically configured to: and taking an Epanechnikov kernel as a kernel function, and intercepting probability density data of a corresponding prediction interval according to the requirements of different confidence degrees.
Preferably, an apparatus, the apparatus comprising: a processor and a memory storing computer program instructions; the processor, when executing the computer program instructions, implements the prediction method of any one of claims 1-6.
Preferably, a computer readable storage medium having computer program instructions stored thereon, which when executed by a processor implement the prediction method of any one of claims 1-6.
The invention provides a short-term load probability prediction method based on a convolutional neural network and quantile regression, which has the beneficial effects that:
1. the method adopts marble loss to replace a loss function of a convolutional neural network, expands a deterministic prediction method into a probabilistic prediction method, and obtains probability density information of each point in a prediction interval at different time points by a kernel density estimation method;
2. the short-term load probability density prediction model based on the convolutional neural network and the quantile regression is a multi-result output model, prediction values under different quantile points can be obtained only through single training, and prediction time is greatly saved.
Drawings
FIG. 1 is a schematic flow chart of a short-term load probability prediction method based on a convolutional neural network and quantile regression.
FIG. 2 is a detailed flowchart of a short-term load probability prediction method based on a convolutional neural network and quantile regression.
FIG. 3 is an hourly load graph of the GEFCom2014 data set.
Fig. 4 is an hourly temperature plot of the GEFCom2014 data set.
Fig. 5 is a graph of load prediction intervals at 80% and 100% confidence on the GEFCom2014 data set based on QRCNN-E.
Detailed Description
In order to make the implementation objects, technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present invention will be described in more detail below with reference to the accompanying drawings in the embodiments of the present invention. In the drawings, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The described embodiments are only some, but not all embodiments of the invention. The embodiments described below with reference to the drawings are illustrative and intended to be illustrative of the invention and are not to be construed as limiting the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The invention constructs a short-term load probability density prediction model (QRCNN-E) based on a convolutional neural network and quantile regression, the model optimizes network parameters by using quantile loss (also called marbles loss) to replace MSE, the convolutional neural network is combined with a quantile regression method, point prediction is expanded into interval prediction, and meanwhile, the defect that the quantile regression method is difficult to solve complex nonlinear problems is overcome. The probability density distribution of the prediction results is then fitted using a kernel density estimation method. Therefore, the power load prediction precision is improved, and high-quality transportation and stable supply of electric energy are guaranteed.
The short-term load probability density prediction model based on the convolutional neural network and the quantile regression uses the following method:
(1) the convolutional neural network method comprises the following steps: the influence of different influence factors (including temperature, electricity price and the like) on the power load is considered, a nonlinear mapping relation is constructed and is subjected to mathematical modeling, and low-dimensional features are abstracted into high-dimensional features.
(2) Quantile regression method: the traditional QR method is one of linear regression methods, can obtain load values under different quantiles at future time by carrying out regression analysis on historical loads and related influence factor data sets, and provides more comprehensive prediction information than point prediction.
(3) The quantile regression convolution neural network method comprises the following steps: the network parameters are optimized by using quantile loss (also called marble loss) to replace MSE, the convolutional neural network is combined with a quantile regression method, the point prediction is expanded into interval prediction, and meanwhile, the defect that the quantile regression method is difficult to solve a complex nonlinear problem is overcome.
(4) The nuclear density estimation method comprises the following steps: and taking the result of the quantile regression convolution neural network as the input of the kernel density estimation method, thereby fitting the probability density distribution of the prediction result. Meanwhile, the influence of different confidence levels on the prediction result is explored.
As shown in fig. 1 and 2, the invention provides a short-term power load probability density prediction method based on a convolutional neural network and quantile regression, which effectively predicts the power load and solves the problem of how to effectively schedule and optimize a power grid system for a power system decision maker.
The method comprises the following steps: collecting power load data, and determining key influence factors according to the relevance of the power load data and external influence factors; we selected 2159 moments of data from 1/2005 to 24/3/2005 as a dataset. Meanwhile, the load value and the temperature of the five moments before, the same moment in the day before, the same moment in the two days before and the same moment in the same day in the week before of the moment to be predicted are selected as input characteristics according to influence factor analysis.
Step two: preprocessing data, supplementing missing values, cleaning abnormal values, normalizing the data to [0,1], and segmenting the data into a training set and a test set;
firstly, supplementing missing data values and cleaning abnormal values by adopting a MinMax method, and then normalizing the data by using a formula (1) to convert a data value domain to [0,1]]Wherein X isnormIs normalized data, x is raw data, xmin、xmaxMaximum and minimum values for data:
the first 80% of the data set was selected as training data and the other 20% as test data. For the quantile regression model, the quantile interval used was 0.01.
Step three: training a QRCNN prediction model, taking quantile loss as a loss function of a convolutional neural network, guiding the network to train, and realizing continuous updating of parameters until the model converges;
the convolutional neural network is combined with a quantile regression method, and the training of a Convolutional Neural Network (CNN) model is guided by substituting quantile loss (also called pinball loss) for mean square error, so that the CNN can predict different quantiles. In order to reduce training cost, marble loss is improved, prediction results under different quantiles are output once through a model, and therefore model efficiency is greatly improved. Meanwhile, the model accuracy is not reduced by verification. The modified pinball loss is shown in formula (2),as a function of the quantile loss function,denotes the j (th)thThe predicted value of the sample point under the i quantile, and q represents the number of quantiles: :
step four: inputting the test data into the trained model to obtain predicted values under different quantiles;
step five: and taking the predicted values under different quantiles as input, and performing load probability density prediction by using a kernel density estimation method under different confidence coefficients to obtain a prediction interval and a probability density curve.
When the nuclear density is estimated, an Epanechnikov nuclear is taken as a nuclear function, and corresponding data are intercepted according to different confidence degree requirements. In experiments with 90% confidence, we cut out quantiles within the [0.1,0.9] interval. At the 100% confidence interval, the full quantile data will be applied.
In another embodiment, the data set of GEFCom2014 is used, and 2159 moments of data from 1/2005 to 3/31/2005 are used as the experimental data set. And simultaneously, selecting the load value and the temperature of the five moments before the moment to be predicted, the same moment in the day before, the same moment in the two days before, and the same moment in the day before the week as input characteristics. For the selection of the temperature information, the invention integrates the temperature information of 25 weather forecast stations, and takes the average value as the temperature at the corresponding moment. The overall distribution of the data set is shown in figures 3 and 4.
As seen from fig. 3 and 4, the electric load has a correlation with the temperature. In particular, the load is higher when the temperature is lower, and the load tends to decrease when the temperature is higher.
On the data set, different kernel density bandwidths are set according to different model quantile regression results so as to obtain the best experimental effect. The bandwidth settings are shown in table 1:
table 1: model bandwidth setting
The results of the point prediction and interval prediction for each model are summarized in table 2 and table 3, respectively:
table 2 point prediction experimental result evaluation based on GEFCom2014 dataset
Table 3 interval prediction experimental result evaluation based on GEFCom2014 dataset
As shown in Table 2, the median value of QRCNN-E gave the best results at 80% confidence level, while the overall performance of QRCNN-E was superior to other methods. It follows that QRCNN-E can extract abstract features that are valuable for prediction by using a convolutional neural network. The experimental result also shows that the deep network has stronger learning capability than the shallow learning model.
Table 3 shows the probabilistic predictions of different approaches on the GEFCom2014 data set. It can be seen that the PICP of each model can obtain coverage values above the confidence level, and QRCNN-E achieves the smallest PINAW at different confidence levels. As can be seen by analyzing the overall evaluation index CWC, the proposed method defeats the remaining comparison algorithms. By combining table 2 and table 3, it can be seen that QRCNN-E not only can achieve a high-precision prediction goal in deterministic prediction, but also can perform best in interval prediction.
The relationship between the upper and lower limits of the prediction interval and the true value at different confidence levels is shown in fig. 5. It turns out that the prediction interval at 100% confidence covers all real values, and the prediction interval at 80% confidence level also covers almost all real values. In addition, the prediction interval at the 80% confidence level is narrower and closer to the actual value than the prediction interval at the 100% confidence level.
According to the power load probability density prediction and evaluation method provided by the invention, the future trend of the power load of the power grid is mastered and the load at the future moment is accurately predicted by analyzing the historical load data and the key influence factors, so that a basis is provided for flexible scheduling of a power system.
The development language of the invention is Python, the development tools are Spyder and Pycharm, and the operating system is Windows 10.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Finally, it should be pointed out that: the above examples are only for illustrating the technical solutions of the present invention, and are not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A short-term power load probability prediction method based on CNN and quantile regression is characterized in that the short-term power load probability density prediction method based on the convolutional neural network and the quantile regression comprises the following steps:
the method comprises the following steps: collecting power load data, and determining input data according to the relevance of the power load data and external influence factors;
step two: preprocessing input data, and dividing the input data into training set data and test set data;
step three: performing short-term load probability density prediction model training based on a convolutional neural network and quantile regression by using the training set data in the step two to obtain a trained short-term load probability density prediction model based on the convolutional neural network and quantile regression;
step four: inputting the test set data in the step two into a trained short-term load probability density prediction model based on a convolutional neural network and quantile regression to obtain predicted values under different quantile points;
step five: and taking the predicted values under different quantiles as input, and performing load probability density prediction by using a kernel density estimation method under different confidence coefficients to obtain a prediction interval and a probability density curve.
2. The CNN and quantile regression-based short-term power load probability prediction method of claim 1, wherein: the step one input data comprises a load value and a temperature.
3. The CNN and quantile regression-based short-term power load probability prediction method of claim 1, wherein: preprocessing input data in the second step, supplementing missing data values and cleaning abnormal values by adopting a MinMax method, and then normalizing the data by using a formula (1) to convert a data value domain to [0,1]]Wherein X isnormIs normalized data, x is raw data, xmin、xmaxMaximum and minimum values for data:
4. the CNN and quantile regression-based short-term power load probability prediction method of claim 1, wherein:
the short-term load probability density prediction model in the third step is obtained by training in the following mode:
and taking the quantile loss as a loss function of the convolutional neural network, and optimizing the parameters of the convolutional neural network to obtain the short-term load probability density prediction model.
5. The CNN and quantile regression-based short-term power load probability prediction method of claim 4, wherein: the quantile loss is shown in equation (2),as a function of the quantile loss function,denotes the j (th)thThe predicted value of the sample point under the i quantile, and q represents the number of quantiles:
6. the CNN and quantile regression-based short-term power load probability prediction method of claim 1, wherein: and the kernel density estimation method of the fifth step adopts an Epanechnikov kernel as a kernel function, and intercepts probability density data of a corresponding prediction interval according to the requirements of different confidence degrees.
7. A short-term power load probability prediction device based on CNN and quantile regression is characterized by comprising the following components:
the acquisition module is used for acquiring power load data and determining input data according to the correlation between the power load data and external influence factors;
the processing module is used for preprocessing input data and dividing the input data into training set data and test set data;
the training module is used for carrying out short-term load probability density prediction model training based on a convolutional neural network and quantile regression by using the training set data to obtain a trained short-term load probability density prediction model based on the convolutional neural network and the quantile regression;
the input module is used for inputting the test set data into a trained short-term load probability density prediction model based on a convolutional neural network and quantile regression to obtain predicted values under different quantile points;
and the prediction module is used for taking the predicted values under different quantiles as input and performing load probability density prediction by using a kernel density estimation method under different confidence coefficients to obtain a prediction interval and a probability density curve.
8. The CNN and quantile regression-based short-term power load probability prediction device of claim 7, wherein:
the prediction module is specifically configured to: and taking an Epanechnikov kernel as a kernel function, and intercepting probability density data of a corresponding prediction interval according to the requirements of different confidence degrees.
9. An apparatus, characterized in that the apparatus comprises: a processor and a memory storing computer program instructions; the processor, when executing the computer program instructions, implements the prediction method of any one of claims 1-6.
10. A computer-readable storage medium having computer program instructions stored thereon which, when executed by a processor, implement the prediction method of any one of claims 1-6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011140762.3A CN112257928A (en) | 2020-10-22 | 2020-10-22 | Short-term power load probability prediction method based on CNN and quantile regression |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011140762.3A CN112257928A (en) | 2020-10-22 | 2020-10-22 | Short-term power load probability prediction method based on CNN and quantile regression |
Publications (1)
Publication Number | Publication Date |
---|---|
CN112257928A true CN112257928A (en) | 2021-01-22 |
Family
ID=74264078
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011140762.3A Pending CN112257928A (en) | 2020-10-22 | 2020-10-22 | Short-term power load probability prediction method based on CNN and quantile regression |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN112257928A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112990536A (en) * | 2021-01-27 | 2021-06-18 | 广东工业大学 | Bus arrival time prediction method based on quantile convolution network |
CN113112092A (en) * | 2021-05-07 | 2021-07-13 | 国网四川省电力公司经济技术研究院 | Short-term probability density load prediction method, device, equipment and storage medium |
CN114498633A (en) * | 2022-02-15 | 2022-05-13 | 国网山东省电力公司菏泽供电公司 | Short-term load probability prediction method based on XGboost and QRLSTM |
CN115409296A (en) * | 2022-11-02 | 2022-11-29 | 国网江西省电力有限公司电力科学研究院 | Method for predicting net load probability of active power distribution network |
CN115841186A (en) * | 2022-12-23 | 2023-03-24 | 国网山东省电力公司东营供电公司 | Industrial park load short-term prediction method based on regression model |
CN117172391A (en) * | 2023-11-02 | 2023-12-05 | 广东电网有限责任公司 | Line loss reasonable interval prediction method, device and medium based on multiple regression |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110648014A (en) * | 2019-08-28 | 2020-01-03 | 山东大学 | Regional wind power prediction method and system based on space-time quantile regression |
CN111080477A (en) * | 2019-12-06 | 2020-04-28 | 北京国电通网络技术有限公司 | Household power load prediction method and system |
US20200143203A1 (en) * | 2018-11-01 | 2020-05-07 | Stephen D. Liang | Method for Design and Optimization of Convolutional Neural Networks |
CN111523924A (en) * | 2020-04-03 | 2020-08-11 | 国家电网有限公司 | Probability prediction method and equipment suitable for electric power spot price |
-
2020
- 2020-10-22 CN CN202011140762.3A patent/CN112257928A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200143203A1 (en) * | 2018-11-01 | 2020-05-07 | Stephen D. Liang | Method for Design and Optimization of Convolutional Neural Networks |
CN110648014A (en) * | 2019-08-28 | 2020-01-03 | 山东大学 | Regional wind power prediction method and system based on space-time quantile regression |
CN111080477A (en) * | 2019-12-06 | 2020-04-28 | 北京国电通网络技术有限公司 | Household power load prediction method and system |
CN111523924A (en) * | 2020-04-03 | 2020-08-11 | 国家电网有限公司 | Probability prediction method and equipment suitable for electric power spot price |
Non-Patent Citations (2)
Title |
---|
许言路等: "基于分位数损失的短期概率负荷预测", 《控制工程》 * |
黄圣权: "基于深度学习的短期电力负荷预测", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112990536A (en) * | 2021-01-27 | 2021-06-18 | 广东工业大学 | Bus arrival time prediction method based on quantile convolution network |
CN112990536B (en) * | 2021-01-27 | 2022-06-03 | 广东工业大学 | Bus arrival time prediction method based on quantile convolution network |
CN113112092A (en) * | 2021-05-07 | 2021-07-13 | 国网四川省电力公司经济技术研究院 | Short-term probability density load prediction method, device, equipment and storage medium |
CN114498633A (en) * | 2022-02-15 | 2022-05-13 | 国网山东省电力公司菏泽供电公司 | Short-term load probability prediction method based on XGboost and QRLSTM |
CN115409296A (en) * | 2022-11-02 | 2022-11-29 | 国网江西省电力有限公司电力科学研究院 | Method for predicting net load probability of active power distribution network |
CN115841186A (en) * | 2022-12-23 | 2023-03-24 | 国网山东省电力公司东营供电公司 | Industrial park load short-term prediction method based on regression model |
CN117172391A (en) * | 2023-11-02 | 2023-12-05 | 广东电网有限责任公司 | Line loss reasonable interval prediction method, device and medium based on multiple regression |
CN117172391B (en) * | 2023-11-02 | 2024-02-27 | 广东电网有限责任公司 | Line loss reasonable interval prediction method, device and medium based on multiple regression |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112257928A (en) | Short-term power load probability prediction method based on CNN and quantile regression | |
CN112488395B (en) | Method and system for predicting line loss of power distribution network | |
CN110163429B (en) | Short-term load prediction method based on similarity day optimization screening | |
CN112990556A (en) | User power consumption prediction method based on Prophet-LSTM model | |
CN110826803A (en) | Electricity price prediction method and device for electric power spot market | |
CN110689190A (en) | Power grid load prediction method and device and related equipment | |
CN116307652A (en) | Artificial intelligent resource allocation method for intelligent power grid | |
CN112396234A (en) | User side load probability prediction method based on time domain convolutional neural network | |
CN112907062A (en) | Power grid electric quantity prediction method, device, medium and terminal integrating temperature characteristics | |
CN111079989A (en) | Water supply company water supply amount prediction device based on DWT-PCA-LSTM | |
CN113361745A (en) | Power distribution network material demand prediction method and system | |
Zhang et al. | Load Prediction Based on Hybrid Model of VMD‐mRMR‐BPNN‐LSSVM | |
CN115358437A (en) | Power supply load prediction method based on convolutional neural network | |
CN108808657B (en) | Short-term prediction method for power load | |
Sun et al. | Short-term power load prediction based on VMD-SG-LSTM | |
CN117277312A (en) | Gray correlation analysis-based power load influence factor method and equipment | |
CN111080037A (en) | Short-term power load prediction method and device based on deep neural network | |
Liu et al. | Embedded Scenario Clustering for Wind and Photovoltaic Power, and Load Based on Multi-Head Self-Attention | |
CN116151464A (en) | Photovoltaic power generation power prediction method, system and storable medium | |
CN115759343A (en) | E-LSTM-based user electric quantity prediction method and device | |
CN115907131A (en) | Method and system for building electric heating load prediction model in northern area | |
Majidpour | Time series prediction for electric vehicle charging load and solar power generation in the context of smart grid | |
Zhu et al. | Prediction of air quality index based on wavelet transform combination model | |
CN110659442B (en) | System, method and device for short-term prediction of data of system, and storage medium | |
Zhang et al. | Load prediction based on depthwise separable convolution model |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210122 |