CN115618710A - Wind power probabilistic prediction method and system based on GAN - Google Patents
Wind power probabilistic prediction method and system based on GAN Download PDFInfo
- Publication number
- CN115618710A CN115618710A CN202211096404.6A CN202211096404A CN115618710A CN 115618710 A CN115618710 A CN 115618710A CN 202211096404 A CN202211096404 A CN 202211096404A CN 115618710 A CN115618710 A CN 115618710A
- Authority
- CN
- China
- Prior art keywords
- wind power
- gan
- data
- generator
- original sample
- 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 51
- 238000013256 Gubra-Amylin NASH model Methods 0.000 claims abstract description 25
- 230000008569 process Effects 0.000 claims abstract description 20
- 238000011156 evaluation Methods 0.000 claims abstract description 14
- 230000000694 effects Effects 0.000 claims abstract description 12
- 238000011478 gradient descent method Methods 0.000 claims abstract description 7
- 238000010276 construction Methods 0.000 claims abstract description 6
- 230000006870 function Effects 0.000 claims description 23
- 238000012549 training Methods 0.000 claims description 23
- 238000004590 computer program Methods 0.000 claims description 14
- 238000009826 distribution Methods 0.000 claims description 10
- 238000003860 storage Methods 0.000 claims description 9
- 238000013527 convolutional neural network Methods 0.000 claims description 4
- 230000007774 longterm Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 5
- 210000002569 neuron Anatomy 0.000 description 4
- 230000004913 activation Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 241000288105 Grus Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- H02J3/004—Generation forecast, e.g. methods or systems for forecasting future energy generation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/08—Probabilistic or stochastic CAD
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/06—Wind turbines or wind farms
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/06—Power analysis or power optimisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
- G06F2119/10—Noise analysis or noise optimisation
-
- 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
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- General Physics & Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Biophysics (AREA)
- Computer Hardware Design (AREA)
- Medical Informatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Geometry (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of wind power prediction, and provides a wind power probabilistic prediction method and a wind power probabilistic prediction system based on a GAN (generic area network), which are used for carrying out probability prediction on wind power based on original sample data of a wind power plant and a trained GAN model to obtain a wind power probability prediction value; wherein, the construction process of the GAN model comprises the following steps: generating false target data according to random noise data with the same dimensionality as original sample data and a generator; on the basis of the false target data, the original sample data of the wind power plant and the discriminator, wassertein distance is introduced to replace JS-KL divergence of original GAN so as to measure the distance between the original sample data of the wind power plant and the false target data, a gradient descent method is adopted to transmit gradient information, parameters of a generator are updated, and a probability prediction value of wind power is obtained according to the fitting result of the two; and calculating to obtain a confidence interval of the predicted value by adopting an evaluation index of the prediction effect based on the wind power probability predicted value.
Description
Technical Field
The invention belongs to the technical field of wind power prediction, and particularly relates to a wind power probabilistic prediction method and system based on GAN (Generative adaptive Networks).
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
With the change of the world energy pattern and the adjustment of the national energy strategy, wind power generation becomes one of the important technical means for promoting the optimization of the energy structure in China. The self uncertainty after the wind power integration has serious influence on the stable operation of the power system. The wind power is accurately predicted, so that the reliability and economy of power grid dispatching can be improved, and the rotating reserve capacity of a conventional unit can be effectively reduced.
The inventor finds that the current probability prediction method for the wind power has the following problems:
in the original GAN, a loss function is diverged based on KL-JS, and in the training process, the GAN model minimizes the difference between two distributions by using cross entropy loss, which is equivalent to minimizing the KL-JS divergence, and the discriminator in the basic GAN is not strong enough to obtain the training result, and meanwhile, the training process is known to be slow and unstable, the time for training the model is also very long, so that good model parameters are difficult to obtain, and an accurate prediction result cannot be obtained.
Disclosure of Invention
In order to solve at least one technical problem in the background art, the invention provides a wind power probabilistic prediction method and system based on GAN, which provides an evaluation index for a prediction effect, calculates a confidence interval of a prediction value, and predicts the future wind power more accurately and intuitively compared with the traditional algorithm.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention provides a wind power probabilistic prediction method based on GAN, which comprises the following steps:
acquiring original sample data of a wind power plant;
performing probability prediction of wind power based on original sample data of the wind power plant and the trained GAN model to obtain a wind power probability prediction value;
wherein, the construction process of the GAN model comprises the following steps:
generating false target data according to random noise data with the same dimensionality as the original sample data and a generator;
on the basis of the false target data, the original sample data of the wind power plant and the discriminator, wassertein distance is introduced to replace JS-KL divergence of original GAN so as to measure the distance between the original sample data of the wind power plant and the false target data, a gradient descent method is adopted to transmit gradient information, parameters of a generator are updated, and a probability prediction value of wind power is obtained according to the fitting result of the two;
and calculating to obtain a confidence interval of the predicted value by adopting an evaluation index of the prediction effect based on the wind power probability predicted value.
A second aspect of the present invention provides a GAN-based wind power probabilistic prediction system, comprising:
the sample acquisition module is used for acquiring original sample data of the wind power plant;
the wind power probability prediction module is used for performing probability prediction on wind power based on original sample data of the wind power plant and the trained GAN model to obtain a wind power probability prediction value;
wherein, the construction process of the GAN model comprises the following steps:
generating false target data according to random noise data with the same dimensionality as original sample data and a generator;
on the basis of the false target data, the original sample data of the wind power plant and the discriminator, wassertein distance is introduced to replace JS-KL divergence of original GAN to measure the distance between the original sample data of the wind power plant and the false target data, a gradient descent method is adopted to transmit gradient information, parameters of a generator are updated, and a probability predicted value of wind power is obtained according to the fitting result of the two;
and the effect evaluation module is used for calculating a confidence interval of the predicted value by adopting an evaluation index of the predicted effect based on the wind power probability predicted value.
A third aspect of the invention provides a computer-readable storage medium.
A computer-readable storage medium, having stored thereon a computer program which, when being executed by a processor, carries out the steps of a GAN-based wind power probabilistic prediction method as described above.
A fourth aspect of the invention provides a computer apparatus.
A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps in a GAN-based wind power probabilistic prediction method as described above when executing the program.
Compared with the prior art, the invention has the beneficial effects that:
the method realizes the wind power prediction regression by carrying out high-precision prediction on the wind power based on the GAN algorithm, the GAN model comprises an identifier, the identifier assists a training generator by generating a countermeasure mode, an evaluation index of a model prediction result is provided, meanwhile, interval prediction is provided, and a confidence interval of the predicted power is provided at a certain confidence level at a certain time in the future. By adopting the prediction method, accurate and reliable probability interval prediction can be obtained, which is not only beneficial to improving the safety of power grid dispatching and effectively reducing the rotation reserve capacity of a conventional unit, but also plays an important role in power transaction.
Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
FIG. 1 is a schematic overall flow chart of a wind power probability prediction method of a GAN in an embodiment of the present invention;
FIG. 2 shows the result of the embodiment of the present invention based on the GAN-GP model training;
FIG. 3 is a graph of the result of the GAN-GP model on the test set after saving the appropriate model parameters according to an embodiment of the present invention;
fig. 4 is an interval prediction diagram illustrating confidence intervals of each segment of a predicted value according to an embodiment of the present invention.
Detailed Description
The invention is further described with reference to the following figures and examples.
It is to be understood that the following detailed description is exemplary and is intended to provide further explanation of the invention as claimed. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the invention. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The training process is unstable as mentioned in the background art; insufficient diversity of the generated samples; the invention solves the problems of collapse of the mode and long training time, and the GAN model of the invention provides Wasserstein distance and gradient punishment to enhance the discriminator, solves the problems of the GAN model, enhances the training stability and strengthens the discriminator to obtain a better generated model.
Example one
As shown in fig. 1, the present embodiment provides a wind power probabilistic prediction method based on GAN, which includes the following steps:
step 1: acquiring original sample data of a wind power plant;
a large amount of historical wind power plant data is needed for training the model, so that the accuracy of model prediction is improved.
In the implementation, a plurality of characteristic data influencing the wind power can be selected and added independently to improve the accuracy of prediction.
Step 2: obtaining technical indexes of original data and wind power trend characteristics from original sample data of a wind power plant;
in one or more embodiments, the calculated technical indicator refers to a moving average of a time series,
the wind power trend characteristics include long-term and short-term trend characteristics of wind power,
wherein the long-term and short-term trends of the wind power are extracted by Fourier transform;
the method specifically comprises the following steps: the fourier transform takes a function and produces a series of sinusoids that, when combined, approximate the original function, which can help the GRU network in the GAN network to more accurately select its predicted trend.
And step 3: the method comprises the steps of obtaining characteristics and a target of original sample data, processing null values and then normalizing the original data.
The target is the target to be predicted, i.e. power.
The characteristics refer to data related to power, such as wind speed and wind direction.
Two parameters are set here: input step size n _ steps _ in and output step size n _ steps _ out,
the number of data we want to enter, and the number of data we want to predict in the future can be set.
The original data set is two-dimensional, and data needs to be reshaped into three-dimensional according to a time step, so that the original data is reshaped into three-dimensional data.
The time step can be set according to real-time requirements, for example, to 30.
And 4, step 4: carrying out probability prediction on wind power according to the remolded three-dimensional original data and the trained GAN model to obtain a wind power probability prediction value;
generating false target data according to random noise data with the same dimensionality as original sample data and a generator;
on the basis of the false target data, the original sample data of the wind power plant and the discriminator, wassertein distance is introduced to replace JS-KL divergence of original GAN to measure the distance between the original sample data of the wind power plant and the false target data, a gradient descent method is adopted to transmit gradient information, parameters of a generator are updated, and a probability predicted value of wind power is obtained according to the fitting result of the two;
in one or more embodiments, in step 4, the generator employs a Gated Round Unit (GRU), and the arbiter employs a Convolutional Neural Network (CNN).
The training process of the GAN model is a process of continuously optimizing the confrontation of a generator and an arbiter, and specifically comprises the following steps:
(1) Constructing structures of a generator and a discriminator;
in the GAN model of this embodiment, the used data set includes a large amount of historical wind farm output power data, and also includes 3 features, which are respectively: wind speed, cosine, sine.
In order to make multi-step look ahead based on raw data, therefore, in the generator, an input step size and an output step size need to be defined, the input of the generator will be a three-dimensional data, i.e., the batch size, the input step size and the characteristics, and the output will be the batch size and the output step size.
To construct a well-behaved generator this embodiment uses three layers of GRUs, with numbers of neurons 1024,512 and 256 respectively, then adds two layers of denses, the last layer of neurons will be the same number as the output step to be predicted.
In the GAN model, the discriminator is a convolutional neural network, which aims to distinguish whether the input data of the discriminator is true or false.
The input to the arbiter is from the raw data and the generated data from the generator. In the discriminator model, it consists of three one-dimensional convolutional layers, 32, 64 and 128 neurons respectively, and three additional sense layers, 220 and 1 neuron respectively, are added at the end.
The activation functions between all layers are set as rectifying linear units (relus), but the output layers do not have Sigmoid activation functions for GAN and linear activation functions for WGAN-GP.
The Sigmoid function will give a single scalar output, 0 and 1, representing true or false.
These two models are merged into the proposed GAN model using the two structures of the generator and the arbiter.
(2) Calculating the loss of the generator and the discriminator by adopting cross entropy;
in the GAN model structure of the present embodiment, cross entropy is used to calculate the loss of the generator and discriminator.
Particularly, in the discriminator, the generated wind power data is combined with the historical power of input steps to be used as the input of the discriminator, so that the data length is increased, and the learning and classifying accuracy of the discriminator is improved.
In the original GAN, the loss function is based on KL-JS divergence, and during training, the GAN model will use cross-entropy loss to minimize the difference between the two distributions, which is equivalent to minimizing the KL-JS divergence.
In this embodiment, the arbiter is trained to maximize its objective function, and the probability of the correct label is assigned to the sample, where the objective function maximized by the arbiter is defined as:
where m is the number of samples, i means the number of samples, y i Mean true samples, x i Is meant as the sample generated.
The generator is then trained to minimize its objective function, i.e.:
wherein, G (x) i ) Is the data (decoy) generated by the generator.
For the calculations presented by the training process in GAN, the penalty function for the arbiter is:
the loss function of the generator is:
by the training process it is always necessary to minimize the loss function to obtain better results.
In step 4, in order to improve the problem that the existing discriminator in GAN is not powerful enough, and the training process is known to be slow and unstable, the present embodiment adopts the Wasserstein distance to solve the problem.
The Wasserstein distance (or Earth Mover Distance (EMD)) is the lowest cost for transmission quality when converting data distribution into data distribution.
The Wasserstein distance of the actual data distribution Pr and the generated data distribution Pg is mathematically defined as the maximum lower bound (infimum) of any transportation plan (i.e. the cost of the cheapest plan):
it should be noted that, while the Wasserstein distance is used to replace the JS-KL divergence of the original GAN, the Wasserstein distance can measure the distance between samples regardless of the situation of the samples, and the current KL-JS divergence cannot be realized.
Since the values of the KL and JS divergence are abrupt, there are only two cases, maximum and minimum, that cannot give useful gradient information, but since the Wassertein distance is not abrupt, and gradient information can be passed using a gradient descent optimization to update the generator's parameters to fit a true distribution.
Where x is the input data to the generator, y is the target from the real dataset, Π (Pr, pg) represents the set of all joint distributions between Pr and Pg, and Π contains all possible transmission plans of γ.
Using the Kantorovich-Rubinstein duality, the calculation can be simplified to:
wherein sup is the minimum upper limit, and f is the 1-Lipschitz function after Lipschitz constraint;
the WGAN-GP enforces the Lipschitz constraint using a gradient penalty.
The differentiable function f is 1-Lipschitz if and only if it has at most 1 everywhere
Gradient of norm. If the gradient norm deviates from its target norm value of 1, the model will be penalized.
And 5: and calculating to obtain a confidence interval of the predicted value by adopting an evaluation index of the prediction effect based on the wind power probability predicted value.
When training is completed, evaluating the performance of the model through Root Mean Square Error (RMSE), and evaluating the total error during prediction, wherein the expected value of the square of the error is represented; the formula is as follows:
the two models are used for confrontation so as to continuously optimize the model parameters, and the accuracy of the models can be ensured. The performance of the model is evaluated through Root Mean Square Error (RMSE), training can be finished when the RMSE reaches a proper range in the training process, and model parameters can be automatically stored.
After the training is finished, a training result is output, as shown in fig. 2, so that the training result can be observed more intuitively.
And substituting the test data set into the stored GAN model, obtaining a prediction result after operation, and outputting a test result, as shown in FIG. 3.
In step 5, the confidence interval of the predicted value is calculated, which uses the standard error and the standard error uses the standard deviation, and they are calculated as follows:
total standard deviation:
sample standard deviation:
standard error:
Assuming that there is a random variable X N (μ, σ) with a mean μ and standard deviation σ, it is expected
If the handle
Standardization of then
Order to
This results in an interval:
where n is the number of samples, x i Is a predicted value, and the method is used,
the mean value of the predicted value is obtained, the obtained wind power prediction data is substituted into a formula to obtain a confidence interval, and the confidence interval shows the degree that the true value has a certain probability to fall around the predicted value.
The results shown in fig. 4 were obtained by programmed calculation. As can be seen from fig. 4, in the implementation, interval prediction is provided through the evaluation index of the model prediction result, and at a certain time in the future, a confidence interval of the predicted power is provided under a certain confidence coefficient.
Example two
The embodiment provides a GAN-based wind power probabilistic prediction system, which includes:
the sample acquisition module is used for acquiring original sample data of the wind power plant;
the wind power probability prediction module is used for performing probability prediction on wind power based on original sample data of the wind power plant and the trained GAN model to obtain a wind power probability prediction value;
wherein, the construction process of the GAN model comprises the following steps:
generating false target data according to random noise data with the same dimensionality as original sample data and a generator;
on the basis of the false target data, the original sample data of the wind power plant and the discriminator, wassertein distance is introduced to replace JS-KL divergence of original GAN so as to measure the distance between the original sample data of the wind power plant and the false target data, a gradient descent method is adopted to transmit gradient information, parameters of a generator are updated, and a probability prediction value of wind power is obtained according to the fitting result of the two;
and the effect evaluation module is used for calculating a confidence interval of the predicted value by adopting an evaluation index of the predicted effect based on the wind power probability predicted value.
EXAMPLE III
The present embodiment provides a computer readable storage medium having stored thereon a computer program which, when being executed by a processor, carries out the steps of a GAN-based wind power probabilistic prediction method as described above.
Example four
The present embodiment provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the GAN-based wind power probabilistic prediction method as described above.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention 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, optical storage, and the like) having computer-usable program code embodied therein.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations 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.
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 a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A wind power probabilistic prediction method based on GAN is characterized by comprising the following steps:
acquiring original sample data of a wind power plant;
performing probability prediction of wind power based on original sample data of the wind power plant and the trained GAN model to obtain a wind power probability prediction value;
wherein, the construction process of the GAN model comprises the following steps:
generating false target data according to random noise data with the same dimensionality as original sample data and a generator;
on the basis of the false target data, the original sample data of the wind power plant and the discriminator, wassertein distance is introduced to replace JS-KL divergence of original GAN so as to measure the distance between the original sample data of the wind power plant and the false target data, a gradient descent method is adopted to transmit gradient information, parameters of a generator are updated, and a probability prediction value of wind power is obtained according to the fitting result of the two;
and calculating to obtain a confidence interval of the predicted value by adopting an evaluation index of the prediction effect based on the wind power probability predicted value.
2. The GAN-based wind power probabilistic prediction method according to claim 1, wherein the technical index of the original sample data of the wind farm is a moving average line of a time series, and the trend characteristics of the wind power comprise long-term and short-term trend characteristics of the wind power.
3. The GAN-based wind power probabilistic prediction method according to claim 1, wherein original sample data of a wind farm is obtained, and then the data is remodeled into three-dimensional data according to a time step.
4. The GAN-based wind power probabilistic prediction method as defined in claim 1, wherein the training process of the GAN model is a process of continuously optimizing the countermeasure of a generator and an arbiter, and specifically comprises:
constructing structures of a generator and a discriminator;
cross entropy is employed to calculate the loss of the generator and the discriminator, where the discriminator is trained to maximize its objective function, the probability of the correct label is assigned to the sample, the generator is trained to minimize its objective function, and training is done with the goal of minimizing the discriminator and the generator minimum loss.
5. The GAN-based wind power probabilistic prediction method as in claim 1, wherein the generator employs a gated-cyclic unit and the arbiter employs a convolutional neural network.
6. The GAN-based wind power probabilistic prediction method according to claim 1, wherein based on Wasserstein distance, kantorovich-Rubinstein duality is adopted for simplification, and the simplified expression is as follows:
wherein, P r Representing actual numbersAccording to the distribution, P g Representing the generated data distribution, x is the input data to the generator, sup is the minimum upper limit, and f is the 1-Lipschitz function after Lipschitz constraints.
7. The GAN-based wind power probabilistic prediction method as defined in claim 1, wherein the evaluation index using the prediction effect includes a total standard deviation, a sample standard deviation, and a standard error.
8. A wind power probabilistic prediction system based on GAN, comprising:
the sample acquisition module is used for acquiring original sample data of the wind power plant;
the wind power probability prediction module is used for performing probability prediction on wind power based on original sample data of the wind power plant and the trained GAN model to obtain a wind power probability prediction value;
wherein, the construction process of the GAN model comprises the following steps:
generating false target data according to random noise data with the same dimensionality as original sample data and a generator;
on the basis of the false target data, the original sample data of the wind power plant and the discriminator, wassertein distance is introduced to replace JS-KL divergence of original GAN so as to measure the distance between the original sample data of the wind power plant and the false target data, a gradient descent method is adopted to transmit gradient information, parameters of a generator are updated, and a probability prediction value of wind power is obtained according to the fitting result of the two;
and the effect evaluation module is used for calculating a confidence interval of the predicted value by adopting an evaluation index of the predicted effect based on the wind power probability predicted value.
9. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of a GAN-based wind power probabilistic prediction method according to any of the claims 1 to 7.
10. A computer arrangement comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor when executing the program carries out the steps in a GAN-based wind power probabilistic prediction method according to any of the claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211096404.6A CN115618710A (en) | 2022-09-08 | 2022-09-08 | Wind power probabilistic prediction method and system based on GAN |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211096404.6A CN115618710A (en) | 2022-09-08 | 2022-09-08 | Wind power probabilistic prediction method and system based on GAN |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115618710A true CN115618710A (en) | 2023-01-17 |
Family
ID=84858342
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211096404.6A Pending CN115618710A (en) | 2022-09-08 | 2022-09-08 | Wind power probabilistic prediction method and system based on GAN |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115618710A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117390967A (en) * | 2023-11-24 | 2024-01-12 | 河海大学 | Offshore wind power short-term power prediction method, device, equipment and storage medium |
| CN117556261A (en) * | 2024-01-08 | 2024-02-13 | 浙江大学 | MCNN-based diaphragm pump check valve service life prediction method and system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111695666A (en) * | 2020-05-26 | 2020-09-22 | 河海大学 | Wind power ultra-short term conditional probability prediction method based on deep learning |
| US20200372297A1 (en) * | 2019-05-23 | 2020-11-26 | Robert Bosch Gmbh | Method for training a generative adversarial network (gan), generative adversarial network, computer program, machine-readable memory medium, and device |
| CN112070209A (en) * | 2020-08-13 | 2020-12-11 | 河北大学 | Stable controllable image generation model training method based on W distance |
| CN114021799A (en) * | 2021-10-29 | 2022-02-08 | 国网辽宁省电力有限公司经济技术研究院 | Day-ahead wind power prediction method and system for wind power plant |
| CN114066236A (en) * | 2021-11-16 | 2022-02-18 | 国家电网公司西南分部 | Wind and solar active power output scene generation method and device, electronic equipment and storage medium |
| US20220197233A1 (en) * | 2020-12-18 | 2022-06-23 | Wuhan University | Wind power prediction method and system for optimizing deep transformer network |
| CN114692950A (en) * | 2022-03-03 | 2022-07-01 | 内蒙古工业大学 | Wind power prediction method |
| CN114861990A (en) * | 2022-04-18 | 2022-08-05 | 云南电网有限责任公司电力科学研究院 | Wind power prediction model modeling method, device and equipment |
-
2022
- 2022-09-08 CN CN202211096404.6A patent/CN115618710A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200372297A1 (en) * | 2019-05-23 | 2020-11-26 | Robert Bosch Gmbh | Method for training a generative adversarial network (gan), generative adversarial network, computer program, machine-readable memory medium, and device |
| CN111695666A (en) * | 2020-05-26 | 2020-09-22 | 河海大学 | Wind power ultra-short term conditional probability prediction method based on deep learning |
| CN112070209A (en) * | 2020-08-13 | 2020-12-11 | 河北大学 | Stable controllable image generation model training method based on W distance |
| US20220197233A1 (en) * | 2020-12-18 | 2022-06-23 | Wuhan University | Wind power prediction method and system for optimizing deep transformer network |
| CN114021799A (en) * | 2021-10-29 | 2022-02-08 | 国网辽宁省电力有限公司经济技术研究院 | Day-ahead wind power prediction method and system for wind power plant |
| CN114066236A (en) * | 2021-11-16 | 2022-02-18 | 国家电网公司西南分部 | Wind and solar active power output scene generation method and device, electronic equipment and storage medium |
| CN114692950A (en) * | 2022-03-03 | 2022-07-01 | 内蒙古工业大学 | Wind power prediction method |
| CN114861990A (en) * | 2022-04-18 | 2022-08-05 | 云南电网有限责任公司电力科学研究院 | Wind power prediction model modeling method, device and equipment |
Non-Patent Citations (2)
| Title |
|---|
| 史洁: "风电场功率超短期预测算法优化研究", 《中国优秀博士论文全文数据库》, 31 December 2012 (2012-12-31) * |
| 花强等: "基于Wasserstein距离的双向学习推理", 《河北大学学报》, vol. 40, no. 3, 25 May 2020 (2020-05-25), pages 328 - 336 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117390967A (en) * | 2023-11-24 | 2024-01-12 | 河海大学 | Offshore wind power short-term power prediction method, device, equipment and storage medium |
| CN117390967B (en) * | 2023-11-24 | 2024-03-29 | 河海大学 | Offshore wind power short-term power prediction method, device, equipment and storage medium |
| CN117556261A (en) * | 2024-01-08 | 2024-02-13 | 浙江大学 | MCNN-based diaphragm pump check valve service life prediction method and system |
| CN117556261B (en) * | 2024-01-08 | 2024-05-14 | 浙江大学 | MCNN-based diaphragm pump check valve service life prediction method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115618710A (en) | Wind power probabilistic prediction method and system based on GAN | |
| CN111929748B (en) | Meteorological element forecasting method and system | |
| Pham et al. | Application of the bees algorithm to the training of learning vector quantisation networks for control chart pattern recognition | |
| US8706670B2 (en) | Relative variable selection system and selection method thereof | |
| CN111382906B (en) | Power load prediction method, system, equipment and computer readable storage medium | |
| Zhang et al. | A novel case adaptation method based on an improved integrated genetic algorithm for power grid wind disaster emergencies | |
| Chinnathambi et al. | Deep neural networks (DNN) for day-ahead electricity price markets | |
| CN109934422A (en) | Neural network wind speed prediction method based on time series data analysis | |
| CN109657884A (en) | Power grid power supply optimization method, apparatus, equipment and computer readable storage medium | |
| CN112241606A (en) | Cooperative decision-making method for operation and maintenance of ship intelligent equipment based on CPS decision-making module | |
| CN114548591B (en) | Sequential data prediction method and system based on mixed deep learning model and Stacking | |
| CN110070228A (en) | BP neural network wind speed prediction method for neuron branch evolution | |
| CN105844353A (en) | Aquatic product price prediction method and device | |
| CN116245019A (en) | Load prediction method, system, device and storage medium based on Bagging sampling and improved random forest algorithm | |
| CN112529053A (en) | Short-term prediction method and system for time sequence data in server | |
| CN115049024A (en) | Training method and device of wind speed prediction model, electronic equipment and storage medium | |
| CN109886488B (en) | Distributed wind power plant layered hybrid short-term prediction method considering wind speed time lag | |
| CN115423091A (en) | Conditional antagonistic neural network training method, scene generation method and system | |
| CN115965119A (en) | Method for power prediction optimization of distributed energy storage system | |
| CN114444763A (en) | Wind power prediction method based on AFSA-GNN | |
| CN116157809A (en) | Method for automatically determining model compression technology based on compression of artificial intelligence model | |
| CN112990274A (en) | Wind power plant abnormal data automatic identification method based on big data | |
| US20200005164A1 (en) | Information processing device, method, program, ic chip and learned model | |
| Chen et al. | Hierarchical fuzzy systems | |
| CN117313555B (en) | GRU humidity prediction method based on self-adaptive OATS-AJSA |
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 |