CN115796351B - Rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation - Google Patents
Rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation Download PDFInfo
- Publication number
- CN115796351B CN115796351B CN202211480610.7A CN202211480610A CN115796351B CN 115796351 B CN115796351 B CN 115796351B CN 202211480610 A CN202211480610 A CN 202211480610A CN 115796351 B CN115796351 B CN 115796351B
- Authority
- CN
- China
- Prior art keywords
- modal
- components
- attenuation
- microwave
- rainfall
- 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.)
- Active
Links
- 238000000354 decomposition reaction Methods 0.000 title claims abstract description 37
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000012545 processing Methods 0.000 claims abstract description 43
- 238000010606 normalization Methods 0.000 claims abstract description 35
- 238000012549 training Methods 0.000 claims abstract description 31
- 238000012360 testing method Methods 0.000 claims description 17
- 238000013178 mathematical model Methods 0.000 claims description 9
- 230000000694 effects Effects 0.000 claims description 7
- 238000007781 pre-processing Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000005457 optimization Methods 0.000 claims description 5
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 claims description 3
- 230000003190 augmentative effect Effects 0.000 claims description 3
- 238000001228 spectrum Methods 0.000 claims description 3
- 230000006870 function Effects 0.000 description 16
- 238000004590 computer program Methods 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000015654 memory Effects 0.000 description 6
- 238000003860 storage Methods 0.000 description 5
- 230000007787 long-term memory Effects 0.000 description 4
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 239000006185 dispersion Substances 0.000 description 1
- 238000013213 extrapolation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
Classifications
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Radar Systems Or Details Thereof (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation. Acquiring a rainless microwave attenuation signal in a sunny day and a microwave attenuation signal in a rainfall period; respectively decomposing the two signals into K modal components; respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components and carrying out normalization processing on the K updated mode components; inputting the processed K modal components in parallel into K LSTM network prediction models obtained by pre-training, and predicting and outputting K modal attenuation components of the next period; reconstructing the K modal attenuation components and performing inverse normalization processing to obtain the attenuation of the microwave signal in the next period, and predicting the attenuation of the microwave signal to obtain the rainfall intensity in the next period. The invention realizes the short-cut rainfall prediction with low cost and high space-time resolution through the variation modal decomposition and the microwave attenuation model.
Description
Technical Field
The invention relates to the technical field of ground meteorological monitoring, in particular to a rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation.
Background
Urban inland inundation disasters are one of the most frequent occurrence and most serious hazard in all natural disasters in China at present. In the face of the outstanding situation, the high-precision rainfall prediction data has remarkable guiding effects on weather forecast, drought and waterlogging disaster early warning and comprehensive treatment. The traditional rainfall prediction method comprises a radar echo extrapolation method and a numerical weather prediction method, but is limited by conventional meteorological observation data such as rainfall station observation, radar observation, satellite observation and the like, so that the cost is high, and the requirements of high space-time resolution and high accuracy are difficult to meet.
Disclosure of Invention
The invention aims to provide a rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation, which can refine and extract different modal characteristics of microwave signal attenuation caused by rainfall, and perform training prediction by utilizing an artificial neural network in different modes, so that the short-term prediction of the rainfall is finally completed, and more accurate rainfall prediction is realized.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
in one aspect, a rainfall short-term prediction method based on variation modal decomposition and microwave attenuation includes:
acquiring a rainless microwave attenuation signal in a sunny day and a microwave attenuation signal in a rainfall period;
decomposing the sunny rainless microwave attenuation signal and the microwave attenuation signal in the rainfall period into K modal components respectively, so that the K modal components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the modal components is minimum; wherein K is a positive integer;
respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components and carrying out normalization processing on the K updated mode components;
inputting the K modal components subjected to normalization treatment into K LSTM network prediction models obtained through pre-training in parallel, and predicting and outputting K modal attenuation components of the next period;
reconstructing the K modal attenuation components and performing inverse normalization processing to obtain the attenuation of the microwave signal of the next period, and predicting and obtaining the rainfall intensity of the next period by utilizing an ITU-R model based on the attenuation of the microwave signal.
Further, the decomposing the sunny non-rain microwave attenuation signal and the rainfall period microwave attenuation signal into K modal components respectively adopts the mathematical model as follows:
wherein u is k 、ω k The method comprises the steps of respectively obtaining the center frequencies of a K-th microwave signal modal component and a K-th modal component of a microwave attenuation signal after corresponding decomposition, wherein K is the total number of modal components obtained by the decomposition, t is time, delta (t) is a dirac function, j is a virtual unit, x is a convolution operator, and is an original microwave signal before the decomposition;
and solving the mathematical model to obtain K modal components.
Further, solving the mathematical model, comprising the steps of:
and introducing a secondary penalty factor and a Lagrange multiplier to convert the mathematical model into an unconstrained variation problem, wherein the expression is as follows:
wherein L (u) k ,ω k Lambda) is an augmented lagrangian expression, lambda is a lagrangian multiplier, and alpha is a secondary penalty factor;
iterative optimization is carried out to obtain each modal component and the center frequency, and an iterative formula is as follows:
wherein n is the iteration number and the initial value is 0; lambda is Lagrangian multiplier, initial value lambda 1 Is 0;respectively the n+1st iteration result of the kth modal component and the center frequency thereof; gamma is noise tolerance; omega is the frequency of the modal spectrum; u (u) i (t) is a modal component other than the kth modal component; />Respectively->u i A frequency domain form of (t), f (t), λ (t);
iterative optimization is continued until the maximum iterative times are reached or the maximum iterative times are satisfied Output the final u k 、ω k Where ε is the preset precision.
Further, pre-training K LSTM network prediction models, including:
acquiring a sample data set, wherein the sample data set comprises a sunny non-rain microwave attenuation signal before rainfall and a plurality of continuous microwave attenuation signals acquired at intervals of the same period during rainfall, and preprocessing the acquired data sample;
decomposing each microwave attenuation signal of the pretreated sunny non-rain microwave attenuation signal and each microwave attenuation signal in the rainfall period into K modal components respectively, so that the K modal components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the modal components is minimum;
the K modal components obtained by decomposing each microwave attenuation signal in the rainfall period are respectively correspondingly subtracted with the modes with the same center frequency in the K modal components obtained by decomposing the rainless microwave attenuation signal in the sunny day, so that a plurality of groups of K updated modal components are obtained;
the K updated modal components are divided into a training set and a testing set after normalization processing;
the method comprises the steps of constructing K LSTM network prediction models, taking K modal components after normalization processing as parallel inputs of the K LSTM network prediction models, taking K modal attenuation components in the next period as outputs of the K LSTM network prediction models, training the K LSTM network prediction models by using a training set, and testing by using a testing set to obtain final K LSTM network prediction models.
Further, the preprocessing the acquired data samples includes: and eliminating the data exceeding the reasonable range and interpolating the lost data.
Further, the loss function used for training is a mean square error loss function, and an Adam optimizer is used as the optimizer.
Further, when the test set is used for testing, the decision coefficient R2 is used for judging the fitting effect, and the formula is as follows:
wherein:predicted values of K modal attenuation components of the next period output by the K LSTM network prediction models; y is i The real values of K modal attenuation components obtained by actually decomposing the microwave attenuation signal of the next period are obtained; />Is y i The average value was calculated.
In another aspect, a rainfall imminent prediction apparatus based on variation modal decomposition and microwave attenuation includes:
the data acquisition module is used for acquiring a rainless microwave attenuation signal in a sunny day and a microwave attenuation signal in a rainfall period;
the variation mode decomposition module is used for respectively decomposing the sunny rainless microwave attenuation signal and the microwave attenuation signal in the rainfall period into K mode components, so that the K mode components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the mode components is minimum; wherein K is a positive integer;
the noise reduction module is used for respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components and carrying out normalization processing on the K updated mode components;
the modal attenuation component prediction module is used for inputting the K modal components subjected to normalization processing into K LSTM network prediction models obtained through pre-training in parallel, and predicting and outputting K modal attenuation components of the next period;
and the rainfall intensity prediction module is used for reconstructing the K modal attenuation components and performing inverse normalization processing to obtain the microwave signal attenuation of the next period, and predicting and obtaining the rainfall intensity of the next period by using the ITU-R model based on the microwave signal attenuation.
Further, the noise reduction module includes:
the first processing module is used for respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components;
and the second processing module is used for carrying out normalization processing on the K updated modal components.
Further, the rainfall intensity prediction module includes:
the third processing module is used for reconstructing the K modal attenuation components and performing inverse normalization processing to obtain predicted microwave signal attenuation;
and the rainfall calculation module is used for calculating and obtaining the rainfall intensity of the next period by utilizing the ITU-R model based on the microwave signal attenuation.
The present invention also provides a computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform the aforementioned method of rainfall short-cut prediction based on variation modal decomposition and microwave attenuation.
The present invention also provides a computing device comprising:
one or more processors, one or more memories, and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, the one or more programs including means for performing the aforementioned variation modal decomposition and microwave attenuation based rainfall short-cut prediction method.
The invention achieves the beneficial technical effects that:
(1) The microwave signals are processed by utilizing the variation modal decomposition, so that the number of modes can be customized, and the time sequence non-stationarity with high complexity and strong nonlinearity can be reduced;
(2) The subtraction of the similar modes of the two types of signals is utilized to finish the noise reduction of the microwave attenuation signals, so that the operation is easier and the stability is better;
(3) And the LSTM network with the corresponding number of modes is utilized to complete prediction, so that different mode characteristics of microwave signal attenuation caused by rainfall can be extracted in a refined manner, and the rainfall prediction with higher accuracy is realized.
(4) The short-term prediction of rainfall is realized by utilizing a microwave attenuation ITU-R model, and the rainfall prediction with more accuracy and higher space-time resolution is realized on the basis of low labor cost and land cost.
Drawings
FIG. 1 is a flow chart of the method of the present invention;
fig. 2 is a schematic view of the structure of the device of the present invention.
Detailed Description
The invention is further described below in connection with specific embodiments. The following examples are only for more clearly illustrating the technical aspects of the present invention, and are not intended to limit the scope of the present invention.
As shown in FIG. 1, the rainfall short-term prediction method based on variation modal decomposition and microwave attenuation comprises the following steps:
step 1, acquiring a rainless microwave attenuation signal in a sunny day and a microwave attenuation signal in a rainfall period;
the non-rain microwave attenuation signal and the microwave attenuation signal in the rainfall period are obtained through a microwave signal transmitter and a receiver. The positions and the number of the microwave signal transmitters and the receivers can be determined according to the spatial resolution required by the target area for rainfall forecast.
Step 2, decomposing the sunny rainless microwave attenuation signal and the microwave attenuation signal in the rainfall period into K modal components respectively, so that the K modal components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the modal components is minimum; wherein K is a positive integer;
step 3, respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components and carrying out normalization processing on the K updated mode components;
step 4, inputting the K modal components subjected to normalization treatment into K LSTM network prediction models obtained through pre-training in parallel, and predicting and outputting K modal attenuation components of the next period in the rainfall period;
and 5, reconstructing K modal attenuation components and performing inverse normalization processing to obtain the attenuation of the microwave signal of the next period, and predicting and obtaining the rainfall intensity of the next period by using an ITU-R model based on the attenuation of the microwave signal.
In the embodiment of the invention, K LSTM prediction models are obtained through training the following steps:
step ss1, acquiring a historical sample data set D1;
the data set D1 comprises microwave attenuation signal intensity data received by the non-rainy microwave signal receiving end in a sunny day and microwave attenuation signal intensity data received by the microwave signal receiving end in a rainy day.
In the embodiment of the present invention, the employed data set D1 includes: a sunny non-rain microwave attenuation signal received 15min before rainfall, and a plurality of continuous microwave attenuation signals received every 15min from the beginning of rainfall during rainfall.
Step ss2, preprocessing the data samples in the data set D1;
wherein the preprocessing comprises the following steps: and removing data exceeding a reasonable range, interpolating lost data and the like.
Step ss3, decomposing the pretreated clear-weather rainless microwave attenuation signals and each microwave attenuation signal in the rainfall period into K modal components respectively, so that the K modal components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the modal components is minimum;
the problem of construction variation of the two types of microwave attenuation signals is solved, the preprocessed two types of microwave signals are respectively decomposed into K modal components, the decomposition sequence is guaranteed to be the modal components with limited bandwidth of the center frequency, meanwhile, the sum of estimated bandwidths of all the modes is minimum, and the constraint condition is that the sum of all the modes is equal to the input signal.
The mathematical model used is as follows:
wherein: u (u) k 、ω k The method comprises the steps of respectively obtaining the center frequencies of a K-th microwave signal modal component and a K-th modal component of a microwave attenuation signal after corresponding decomposition, wherein K is the total number of modal components obtained through decomposition, delta (t) is a dirac function, j is a virtual unit, x is a convolution operator, and f is an original microwave signal before decomposition;
the second order penalty factor alpha and Lagrange multiplier lambda are introduced to convert the constrained variation problem into the unconstrained variation problem, and the expression is as follows:
wherein L (u) k ,ω k Lambda) is an augmented lagrangian expression;
iterative optimization is carried out to obtain each modal component and the center frequency, and an iterative formula is as follows:
wherein, the iteration times are the initial value of 0; lambda is Lagrangian multiplier, initial value lambda 1 Is 0;respectively the n+1st iteration result of the kth modal component and the center frequency thereof; gamma is noise tolerance; omega is the frequency of the modal spectrum; u (u) i (t) is a modal component other than the kth modal component; />Respectively->u i (t), f (t), λ ();
when the iteration is optimized until the maximum iteration number is reached or the maximum iteration number is satisfied Outputting the final u k 、ω k 。
Wherein: epsilon is a precision convergence criterion.
Step ss4, correspondingly subtracting the K modal components obtained by decomposing each microwave attenuation signal in the rainfall period from the modes with the same center frequency in the K modal components obtained by decomposing the rainless microwave attenuation signal in the sunny day to obtain a plurality of groups of updated K modal components, and constructing the K modal components as a data set D2;
and correspondingly subtracting the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals, so that the microwave attenuation caused by non-rainfall can be eliminated, and the denoising effect is achieved.
In the data set D2, each data sample includes a set of updated K modal components.
Step ss5, carrying out normalization processing on each sample in the data set D2, dividing the sample into a training set and a testing set, constructing K LSTM network prediction models, setting corresponding parameters, and training and testing the model;
the samples in the dataset D2 are subjected to a dispersion normalization process, the formula is as follows:
from the processed dataset, 70% was evenly extracted as training set and 30% as test set.
K LSTM prediction models are correspondingly constructed for K modal components based on Tensorflow 2.0 environment. Each LSTM prediction model includes an input layer, a hidden layer, and an output layer. The prediction purpose is achieved by continuously updating long-term memory and screening.
The hidden layer of the model uses a forgetting gate (forget gate) in forward calculation, and decides whether the information in long-term memory needs to be forgotten or not according to the information in short-term memory and the current input information; an input gate (input gate) for updating information in the long-term memory in combination with a sigmoid function and a tanh function; and an output gate (output gate), screening effective information in the long-term memory by using a sigmoid function, and combining the effective information with a tanh function to finish final output. The formula is as follows:
forgetting the door: f (f) t ＝σ(W f ·[h t-1 ,x t ]+ f )
An input door: i.e t ＝σ(W i ·[h t-1 ,x t ]+ i )
Output door: o (o) t ＝σ(W o [h t-1 ,x t ]+b o )
Wherein: sigma is a sigmoid function; tan h is the t hyperbolic tangent function; x is x t Representing a sample input at the current time; h is a t-1 Is the output of the previous moment; w (W) f 、W c 、W f 、W o Is the weight; b f 、b c 、b i 、b o Is biased.
Wherein, the loss function uses a mean square error loss function, and the formula is as follows:
wherein: n is the number of samples, namely, the batch size;is a predicted value; y is the actual value.
The optimizer uses an Adam optimizer with which the LSTM predictive network model described above is updated continuously. The super parameters are set as follows:
1. learning rate l r ＝0.001
2. Rate of decay beta 1 ＝0.9，β 2 ＝0.999
And taking the K modal components as parallel inputs of the K LSTM network prediction models, taking the K modal attenuation components of the next period as outputs of the K LSTM network prediction models, and training the K LSTM network prediction models by using a training set.
After training, inputting the test sample into K LSTM network prediction models, and judging the fitting effect by using the decision coefficient R2, wherein the formula is as follows.
Wherein:predicted values of K modal attenuation components of the next period output by the K LSTM network prediction models; y is i The real values of K modal attenuation components obtained by actually decomposing the microwave attenuation signal of the next period are obtained; />Is y i The average value was calculated.
Judging whether the model is good or bad according to the value of R2, wherein the value range is [0,1], the larger the R2 is, the better the fitting effect is, 0.9 is selected as a threshold, and the training is completed after the threshold is exceeded.
And correspondingly predicting K modal attenuation components by using K LSTM network prediction models after training, carrying out inverse normalization after reconstruction, and finishing short-term rainfall prediction according to the ITU-R model.
In step 5, reconstructing K modal attenuation components, including: and adding K modal attenuation components predicted by the model to obtain predicted microwave signal attenuation.
Then, based on the predicted attenuation of the microwave signal, the rainfall intensity is predicted by using an ITU-R model, and the short-term prediction of rainfall is completed, wherein the calculation formula is as follows:
R＝ak b
wherein: r is the rainfall intensity of the next period of the predicted rainfall period; k is the microwave attenuation of the next period of the rainfall period predicted; a. b are coefficients, respectively.
As shown in fig. 2, a rainfall short-term prediction device based on variation modal decomposition and microwave attenuation includes:
the data acquisition module is used for acquiring a rainless microwave attenuation signal in a sunny day and a microwave attenuation signal in a rainfall period;
the variation mode decomposition module is used for respectively decomposing the sunny rainless microwave attenuation signal and the microwave attenuation signal in the rainfall period into K mode components, so that the K mode components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the mode components is minimum; wherein K is a positive integer;
the noise reduction module is used for respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components and carrying out normalization processing on the K updated mode components;
the modal attenuation component prediction module is used for inputting the K modal components subjected to normalization processing into K LSTM network prediction models obtained through pre-training in parallel, and predicting and outputting K modal attenuation components of the next period;
and the rainfall intensity prediction module is used for reconstructing the K modal attenuation components and performing inverse normalization processing to obtain the microwave signal attenuation of the next period, and predicting and obtaining the rainfall intensity of the next period by using the ITU-R model based on the microwave signal attenuation.
As shown in fig. 2, the noise reduction module includes:
the first processing module is used for respectively carrying out corresponding subtraction on the K modal components obtained by decomposing the two types of signals and the modes with the same center frequency to obtain K updated modal components;
and the second processing module is used for carrying out normalization processing on the K updated modal components.
As shown in fig. 2, the rainfall intensity prediction module includes:
the third processing module is used for reconstructing the K modal attenuation components and performing inverse normalization processing to obtain predicted microwave signal attenuation;
and the rainfall calculation module is used for calculating and obtaining the rainfall intensity of the next period by utilizing the ITU-R model based on the microwave signal attenuation.
The present invention also provides a computer readable storage medium storing one or more programs, the one or more programs comprising instructions, which when executed by a computing device, cause the computing device to perform the aforementioned method of rainfall short-cut prediction based on variation modal decomposition and microwave attenuation.
The present invention also provides a computing device comprising:
one or more processors, one or more memories, and one or more programs, wherein the one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, the one or more programs including means for performing the aforementioned variation modal decomposition and microwave attenuation based rainfall short-cut prediction method.
It will be appreciated by those skilled in the art that 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 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.
The foregoing is merely a preferred embodiment of the present invention, and it should be noted that modifications and variations could be made by those skilled in the art without departing from the technical principles of the present invention, and such modifications and variations should also be regarded as being within the scope of the invention.
Claims (9)
1. The rainfall short-term prediction method based on variation modal decomposition and microwave attenuation is characterized by comprising the following steps of:
acquiring a rainless microwave attenuation signal in a sunny day and a microwave attenuation signal in a rainfall period;
decomposing the sunny rainless microwave attenuation signal and the microwave attenuation signal in the rainfall period into K modal components respectively, so that the K modal components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the modal components is minimum; wherein K is a positive integer;
respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components and carrying out normalization processing on the K updated mode components;
inputting the K modal components subjected to normalization treatment into K LSTM network prediction models obtained through pre-training in parallel, and predicting and outputting K modal attenuation components of the next period;
reconstructing K modal attenuation components and performing inverse normalization processing to obtain microwave signal attenuation of the next period, and predicting rainfall intensity of the next period by using an ITU-R model based on the microwave signal attenuation;
pre-training K LSTM network prediction models, comprising:
acquiring a sample data set, wherein the sample data set comprises a sunny non-rain microwave attenuation signal before rainfall and a plurality of continuous microwave attenuation signals acquired at intervals of the same period during rainfall, and preprocessing the acquired data sample;
decomposing each microwave attenuation signal of the pretreated sunny non-rain microwave attenuation signal and each microwave attenuation signal in the rainfall period into K modal components respectively, so that the K modal components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the modal components is minimum;
the K modal components obtained by decomposing each microwave attenuation signal in the rainfall period are respectively correspondingly subtracted with the modes with the same center frequency in the K modal components obtained by decomposing the rainless microwave attenuation signal in the sunny day, so that a plurality of groups of K updated modal components are obtained;
the K updated modal components are divided into a training set and a testing set after normalization processing;
the method comprises the steps of constructing K LSTM network prediction models, taking K modal components after normalization processing as parallel inputs of the K LSTM network prediction models, taking K modal attenuation components in the next period as outputs of the K LSTM network prediction models, training the K LSTM network prediction models by using a training set, and testing by using a testing set to obtain final K LSTM network prediction models.
2. The rainfall short-term prediction method based on variation modal decomposition and microwave attenuation according to claim 1, wherein the decomposing of the sunny and rainless microwave attenuation signal and the microwave attenuation signal during rainfall into K modal components respectively adopts the mathematical model:
in the method, in the process of the invention,u k 、ω k the method comprises the steps of respectively obtaining the center frequencies of a K-th microwave signal modal component and a K-th modal component of a microwave attenuation signal after corresponding decomposition, wherein K is the total number of modal components obtained through decomposition, t is time, delta (t) is a dirac function, j is a virtual unit, x is a convolution operator, and f is an original microwave signal before decomposition;
and solving the mathematical model to obtain K modal components.
3. The rainfall short-term prediction method based on variation modal decomposition and microwave attenuation according to claim 2, wherein solving the mathematical model comprises the steps of:
and introducing a secondary penalty factor and a Lagrange multiplier to convert the mathematical model into an unconstrained variation problem, wherein the expression is as follows:
wherein L (u) k ,ω k Lambda) is an augmented lagrangian expression, lambda is a lagrangian multiplier, and alpha is a secondary penalty factor;
iterative optimization is carried out to obtain each modal component and the center frequency, and an iterative formula is as follows:
wherein n is the iteration number and the initial value is 0; lambda is Lagrangian multiplier, initial value lambda 1 Is 0;respectively the n+1st iteration result of the kth modal component and the center frequency thereof; gamma is noise tolerance; omega is the frequency of the modal spectrum; u (u) i (t) is a modal component other than the kth modal component; />Respectively isu i A frequency domain form of (t), f (t), λ (t);
iterative optimization is continued until the maximum iterative times are reached or the maximum iterative times are satisfied Output the final u k 、ω k Where ε is the preset precision.
4. The method for short-term rainfall prediction based on variation modal decomposition and microwave attenuation according to claim 1, wherein the preprocessing of the acquired data samples comprises: and eliminating the data exceeding the reasonable range and interpolating the lost data.
5. The rainfall short-cut prediction method based on variation modal decomposition and microwave attenuation according to claim 1, wherein the loss function used in training is a mean square error loss function, and an Adam optimizer is used in the optimizer.
6. The rainfall short-term prediction method based on variation modal decomposition and microwave attenuation according to claim 1, wherein the fitting effect is judged by using a decision coefficient R2 when a test set is used for testing, and the formula is as follows:
wherein:predicted values of K modal attenuation components of the next period output by the K LSTM network prediction models; y is i The real values of K modal attenuation components obtained by actually decomposing the microwave attenuation signal of the next period are obtained; />Is y i The average value was calculated.
7. Rainfall short-term prediction device based on variation modal decomposition and microwave attenuation is characterized by comprising:
the data acquisition module is used for acquiring a rainless microwave attenuation signal in a sunny day and a microwave attenuation signal in a rainfall period;
the variation mode decomposition module is used for respectively decomposing the sunny rainless microwave attenuation signal and the microwave attenuation signal in the rainfall period into K mode components, so that the K mode components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the mode components is minimum; wherein K is a positive integer;
the noise reduction module is used for respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components and carrying out normalization processing on the K updated mode components;
the modal attenuation component prediction module is used for inputting the K modal components subjected to normalization processing into K LSTM network prediction models obtained through pre-training in parallel, and predicting and outputting K modal attenuation components of the next period;
the rainfall intensity prediction module is used for reconstructing K modal attenuation components and performing inverse normalization processing to obtain microwave signal attenuation of the next period, and based on the microwave signal attenuation, rainfall intensity of the next period is predicted by using an ITU-R model;
wherein, training K LSTM network prediction models in advance includes:
acquiring a sample data set, wherein the sample data set comprises a sunny non-rain microwave attenuation signal before rainfall and a plurality of continuous microwave attenuation signals acquired at intervals of the same period during rainfall, and preprocessing the acquired data sample;
decomposing each microwave attenuation signal of the pretreated sunny non-rain microwave attenuation signal and each microwave attenuation signal in the rainfall period into K modal components respectively, so that the K modal components meet the limited bandwidth with the center frequency, and the sum of the estimated bandwidths of the modal components is minimum;
the K modal components obtained by decomposing each microwave attenuation signal in the rainfall period are respectively correspondingly subtracted with the modes with the same center frequency in the K modal components obtained by decomposing the rainless microwave attenuation signal in the sunny day, so that a plurality of groups of K updated modal components are obtained;
the K updated modal components are divided into a training set and a testing set after normalization processing;
the method comprises the steps of constructing K LSTM network prediction models, taking K modal components after normalization processing as parallel inputs of the K LSTM network prediction models, taking K modal attenuation components in the next period as outputs of the K LSTM network prediction models, training the K LSTM network prediction models by using a training set, and testing by using a testing set to obtain final K LSTM network prediction models.
8. The rainfall short-term prediction device based on variation modal decomposition and microwave attenuation according to claim 7, wherein the noise reduction module comprises:
the first processing module is used for respectively carrying out corresponding subtraction on the modes with the same center frequency in the K mode components obtained by decomposing the two types of signals to obtain K updated mode components;
and the second processing module is used for carrying out normalization processing on the K updated modal components.
9. The variation modal decomposition and microwave attenuation based rainfall imminent prediction device of claim 7, wherein the rainfall intensity prediction module comprises:
the third processing module is used for reconstructing the K modal attenuation components and performing inverse normalization processing to obtain predicted microwave signal attenuation;
and the rainfall calculation module is used for calculating and obtaining the rainfall intensity of the next period by utilizing the ITU-R model based on the microwave signal attenuation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211480610.7A CN115796351B (en) | 2022-11-24 | 2022-11-24 | Rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211480610.7A CN115796351B (en) | 2022-11-24 | 2022-11-24 | Rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115796351A CN115796351A (en) | 2023-03-14 |
| CN115796351B true CN115796351B (en) | 2024-03-22 |
Family
ID=85440905
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211480610.7A Active CN115796351B (en) | 2022-11-24 | 2022-11-24 | Rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115796351B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116485010B (en) * | 2023-03-20 | 2024-04-16 | 四川省雅安市气象局 | S2S precipitation prediction method based on cyclic neural network |
| CN116383701B (en) * | 2023-04-13 | 2024-01-26 | 中国人民解放军国防科技大学 | Microwave link rainy period detection method based on learning reconstruction |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20170097249A (en) * | 2016-02-17 | 2017-08-28 | 부경대학교 산학협력단 | Rainfall Estimatation System using Microwave link and Method of the same |
| KR101915665B1 (en) * | 2018-01-22 | 2018-11-06 | 부경대학교 산학협력단 | Monitoring system for rainy and monitoring method using same by microwave attenuation character due to rainfall |
| CN109936855A (en) * | 2017-12-18 | 2019-06-25 | 中国移动通信集团海南有限公司 | Method, apparatus, equipment and the medium being distributed using base station signal retrieving precipitation |
| WO2021197167A1 (en) * | 2020-03-30 | 2021-10-07 | 国防科技大学 | Rainfall and water vapor comprehensive measurement device and method based on dual-frequency dual-polarization microwave link |
| CN113705864A (en) * | 2021-08-13 | 2021-11-26 | 华北水利水电大学 | Weather drought prediction method and device based on VMD-CNN-BilSTM-ATT mixed model |
| CN114036850A (en) * | 2021-11-16 | 2022-02-11 | 南昌工程学院 | Runoff prediction method based on VECGM |
-
2022
- 2022-11-24 CN CN202211480610.7A patent/CN115796351B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20170097249A (en) * | 2016-02-17 | 2017-08-28 | 부경대학교 산학협력단 | Rainfall Estimatation System using Microwave link and Method of the same |
| CN109936855A (en) * | 2017-12-18 | 2019-06-25 | 中国移动通信集团海南有限公司 | Method, apparatus, equipment and the medium being distributed using base station signal retrieving precipitation |
| KR101915665B1 (en) * | 2018-01-22 | 2018-11-06 | 부경대학교 산학협력단 | Monitoring system for rainy and monitoring method using same by microwave attenuation character due to rainfall |
| WO2021197167A1 (en) * | 2020-03-30 | 2021-10-07 | 国防科技大学 | Rainfall and water vapor comprehensive measurement device and method based on dual-frequency dual-polarization microwave link |
| CN113705864A (en) * | 2021-08-13 | 2021-11-26 | 华北水利水电大学 | Weather drought prediction method and device based on VMD-CNN-BilSTM-ATT mixed model |
| CN114036850A (en) * | 2021-11-16 | 2022-02-11 | 南昌工程学院 | Runoff prediction method based on VECGM |
Non-Patent Citations (3)
| Title |
|---|
| 基于微波链路的路径雨强反演方法及实验研究;高太长 等;物理学报;64(7);174301-1~174301-9 * |
| 基于时频分析的LSTM组合模型径流预测;蔡文静 等;干旱区地理;第44卷(第6期);1696-1706 * |
| 微波通信链路监测降水试验及可行性探究;高清泉;韩;肖天贵;;成都信息工程大学学报(02);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115796351A (en) | 2023-03-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115796351B (en) | Rainfall short-term prediction method and device based on variation modal decomposition and microwave attenuation | |
| CN109886498B (en) | EMD-GRU short-term power load prediction method based on feature selection | |
| CN110135580B (en) | Convolution network full integer quantization method and application method thereof | |
| CN110334875B (en) | Wind power combination probability prediction method considering evaluation index conflict | |
| CN107092744B (en) | Surface subsidence prediction method based on EMD-SVR | |
| CN110309603B (en) | Short-term wind speed prediction method and system based on wind speed characteristics | |
| CN113687433B (en) | Bi-LSTM-based magnetotelluric signal denoising method and system | |
| CN115951430B (en) | Geomagnetic index forecasting method based on LSTM | |
| CN115587666A (en) | Load prediction method and system based on seasonal trend decomposition and hybrid neural network | |
| CN116451553A (en) | Improved variation modal decomposition and BiGRU fusion water quality prediction method | |
| CN112101524A (en) | Method and system for on-line switching bit width quantization neural network | |
| CN112418476A (en) | Ultra-short-term power load prediction method | |
| CN116453096A (en) | Image foreign matter detection method, device, electronic equipment and storage medium | |
| CN113836823A (en) | Load combination prediction method based on load decomposition and optimized bidirectional long-short term memory network | |
| CN110807428B (en) | Coal sample identification method, device, server and storage medium | |
| CN115423145A (en) | Photovoltaic power prediction method based on Kmeans-VMD-WT-LSTM method | |
| CN105303051A (en) | Air pollutant concentration prediction method | |
| CN114500335A (en) | SDN network flow control method based on fuzzy C-means and mixed kernel least square support vector machine | |
| CN113449920A (en) | Wind power prediction method, system and computer readable medium | |
| CN116701875A (en) | Audible noise probability prediction method and system for extra-high voltage alternating current transmission line | |
| CN116189008A (en) | Remote sensing image change detection method based on fixed point number quantification | |
| CN111104640A (en) | Rainfall observation and evaluation method and system based on analytic hierarchy process | |
| CN116186498A (en) | Earthquake signal noise suppression method and system based on self-supervision learning | |
| CN112287607B (en) | Multi-scale ionized layer TEC prediction method | |
| CN115860165A (en) | Neural network basin rainfall runoff forecasting method and system considering initial loss |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |