CN113033094B - Sea wave height prediction method - Google Patents
Sea wave height prediction method Download PDFInfo
- Publication number
- CN113033094B CN113033094B CN202110313922.8A CN202110313922A CN113033094B CN 113033094 B CN113033094 B CN 113033094B CN 202110313922 A CN202110313922 A CN 202110313922A CN 113033094 B CN113033094 B CN 113033094B
- Authority
- CN
- China
- Prior art keywords
- wave height
- pimf
- sea wave
- model
- original
- 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
- 238000000034 method Methods 0.000 title claims abstract description 20
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 14
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000006403 short-term memory Effects 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 4
- 230000015654 memory Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 241000282414 Homo sapiens Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/08—Fluids
-
- 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
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Hardware Design (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (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)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a sea wave height prediction method, which comprises the following steps: performing VMD decomposition on the original sea wave height data sequence; decomposing the VMD to obtain a plurality of discrete subsequences IMF 1 ，IMF 2 ……IMF k Inputting the result into an AM-LSTM model to obtain a model prediction result PIMF of the discrete subsequences 1 ，PIMF 2 ……PIMF k The method comprises the steps of carrying out a first treatment on the surface of the Inputting the original sea wave height data into the AM-LSTM model to obtain a model prediction result PIMF of the original sea wave height data sequence; model predictive result PIMF for the plurality of discrete subsequences 1 ，PIMF 2 ……PIMF k Reconstructing and calculating the PIMF (model predictive model) of the original sea wave height data sequence to obtain the final predicted value of the sea wave heightAnd evaluating the final predicted value.
Description
Technical Field
The invention relates to the field of time sequence prediction, in particular to a sea wave height prediction method based on variation modal decomposition and a attention mechanism.
Background
Sea waves are a complex three-dimensional random motion and are one of the very important power processes in the ocean. The height of sea waves is an important factor affecting the related activities of human beings such as offshore operations, coastal resident lives, harbor and wharfs construction and the like. Many scholars at home and abroad have conducted intensive research on the prediction of sea wave height. The traditional numerical prediction model is based on an energy balance equation, and some traditional numerical wave models are used for predicting sea wave height, however, due to factors such as various inputs of a large amount of data, complexity of calculation, boundary conditions and the like, the calculation of a large model has very severe requirements on the performance requirement and the time cost of a computer. Early studies of sea wave altitude forecasting used classical time series models, however, classical time series models are not suitable for predicting non-linear and non-stationary data due to their linear and stationary assumptions.
In recent years, many scholars have begun to conduct wave height prediction studies using soft computing methods, for example using a hybrid model based on EMD (Empirical Mode Decomposition ) and SVR (SupportVector Regression, support vector regression) for short-term prediction of nonlinear non-stationary wave heights. EMD is a data-driven decomposition technique that is adaptive in analyzing non-linear and non-stationary data sets, however, the inherent modal aliasing in empirical mode decomposition is a disadvantage that compromises the performance of empirical mode decomposition. Therefore, how to solve the problem that the sea wave height prediction accuracy is not high due to the nonlinearity and the non-stationarity of the sea wave height is a problem to be solved at present.
Disclosure of Invention
The invention aims to solve the technical problem of low sea wave height prediction precision caused by nonlinearity and non-stationarity of sea wave height, and provides a sea wave height prediction method.
The invention solves the technical problems by the following technical scheme:
a method of predicting sea wave height, the method comprising:
performing VMD decomposition on the original sea wave height data sequence;
decomposing the VMD to obtain a plurality of discrete subsequences IMF 1 ，IMF 2 ……IMF k Inputting the result into an AM-LSTM model to obtain a model prediction result PIMF of the discrete subsequences 1 ，PIMF 2 ……PIMF k ；
Inputting the original sea wave height data into the AM-LSTM model to obtain a model prediction result PIMF of the original sea wave height data sequence;
model predictive result PIMF for the plurality of discrete subsequences 1 ，PIMF 2 ……PIMF k Reconstructing and calculating the PIMF (model predictive model) of the original sea wave height data sequence to obtain the final predicted value of the sea wave height
And evaluating the final predicted value.
Preferably, the VMD decomposition of the original sea wave height data sequence comprises:
initializing parameters of the VMDSetting the maximum iteration number n;
for the saidIteration is performed according to the following formula:
wherein lambda is Lagrangian multiplier, alpha is a quadratic penalty factor, tau is noise tolerance, i is imaginary number, k is the number of the plurality of discrete subsequences, w is center frequency, u k ，w k Respectively the kth discrete subsequencesAnd the center frequency of the kth discrete sub-sequence,fourier transforms of f (t), values obtained by a multiplication operator alternating direction algorithm;
the iteration is carried out until the conditionUp to the point where the precision converges on criterion epsilon>0；
Output the final u k ，w k 。
Further, the reconstruction calculation obtains the final predicted value of the sea wave heightComprising the following steps:
further, the final predicted value is evaluated using the mean absolute error, root mean square error, and mean absolute percentage error.
On the basis of conforming to the common knowledge in the field, the above preferred conditions can be arbitrarily combined to obtain the preferred examples of the invention.
The invention has the positive progress effects that: the original sea wave height data is decomposed into a plurality of more stable subsequences through the VMD, so that the original sea wave height data is easier to predict by a machine learning method, and the prediction precision is improved; the LSTM model based on the attention mechanism is creatively applied to subsequence prediction, the predicted data are reconstructed, and the final prediction result is compared with the true value, so that the prediction result is more approximate to the true value; the proposed hybrid model not only remarkably improves the prediction precision of sea wave height, but also has stronger generalization performance, and is suitable for the prediction of sea wave height.
Drawings
FIG. 1 is a flow chart of a method for predicting sea wave height according to an embodiment of the present invention.
Detailed Description
In order to facilitate an understanding of the present application, a more complete description of the present application will now be provided with reference to the relevant figures. Preferred embodiments of the present application are shown in the drawings. This application may, however, be embodied in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
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 application belongs. The terminology used herein in the description of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The term "and/or" as used herein includes any and all combinations of one or more of the associated listed items.
A method flow diagram in one embodiment of the invention is shown in fig. 1:
in one example, we selected the sea area site of gulfweed in the eastern united states for sea wave height prediction, all within acceptable tolerances specified by the national buoy center, while meeting the relevant specifications for WMO (World Meteorological Organization ).
The first step: decomposing original sea wave height data of a gulfweed sea area station into 6 subsequences IMF by adopting variation modal decomposition of a signal processing technology 1 、IMF 2 、IMF 3 、IMF 4 、IMF 5 、IMF 6 (the number of decompositions per site is not necessarily the same), decomposing the non-linear non-stationary wave height time series data into smoother and more regular sub-sequences. Sea wave height data is recorded every other hour for a total of 4392 data from day 1, 6, 2016 to day 11, 30, 2016.
The method comprises the steps of decomposing original sea wave height data of a gulfweed sea area station by using the following iterative relationship:
wherein lambda is Lagrangian multiplier, alpha is a quadratic penalty factor, tau is noise tolerance, i is imaginary number, k is the number of the plurality of discrete subsequences, w is center frequency, u k ，w k The kth discrete sub-sequence and the center frequency of said kth discrete sub-sequence,fourier transforms, which are f (t), are iterated until ++is satisfied by the value obtained by the multiplication operator alternating direction algorithm>Up to the point where the precision converges on criterion epsilon>0, output the final u k ，w k 。
And a second step of: the LSTM (Long Short Term Memory, long-short-term memory artificial neural network) model can be further improved by focusing the input time series data differently by using the attention mechanism, and an AM-LSTM (Attentional Mechanism-Long Short Term Memory, long-short-term memory artificial neural network) model is formed. Predicting the original sea wave height data through an AM-LSTM model to obtain a predicted sequence PIMF; the 6 subsequences obtained in the first step are respectively predicted by an AM-LSTM model, and the corresponding predicted sequence is PIMF 1 、PIMF 2 、PIMF 3 、PIMF 4 、PIMF 5 、PIMF6。
And a third step of: reconstructing the result PIMF of the original wave height data through model prediction and the result of each subsequence through model prediction. The reconstruction mode is shown as follows:
and obtaining a final prediction result after reconstruction.
Fourth step: the results obtained after reconstruction were evaluated against the true values of the original data using MAE (Mean Absolute Error ), RMSE (RootMean Square Error, root mean square error), MAPE (Mean Absolute Percentage Error ). Meanwhile, the model is compared with LSTM, VMD-LSTM (Variational Mode Decomposition-Long ShortTerm Memory, variable modal decomposition-long-short-term memory artificial neural network) and AM-LSTM and other similar models. The comparison results are shown in Table 1.
TABLE 1
Table 1 lists the conditions of eight algorithms in different evaluation indexes after training and testing, and because the sea wave height data has nonlinearity and non-stationarity, the sea wave height data is decomposed into relatively stable subsequences with different frequency scales by adopting VMD (Variational ModeDecomposition, variational modal decomposition), and hidden information in the sea wave height data is deeply mined. The prediction accuracy after VMD decomposition of the original data is higher than that of the original data directly. The LSTM model incorporating the attention mechanism is able to selectively pay attention to the time-series data of the input and to the data carrying more information, so its accuracy of prediction is generally superior to that of LSTM without the attention mechanism. By employing a reconstruction method such that its predictive performance is higher than a model without the reconstruction method, it is meaningful to explain the manner of reconstruction.
By comparing different models, according to three evaluation indexes of MAE, RMSE and MAE, the VALM (Variational mode decomposition Attentional mechanism Long short term Memory, variable mode attention mechanism long-term and short-term memory artificial neural network) hybrid model can more effectively predict sea wave height data, and incomparable performance of a single model is generated, and meanwhile, the prediction accuracy of the hybrid model is superior to that of other hybrid models. It can be seen that compared with other algorithms, the error of MAE of 0.1016, RMSE of 0.1424 and MAPE of 5.9913 is lower than that of other algorithms when the wave height is predicted by the VALM algorithm. Based on these findings, we can conclude that: compared with other single models or mixed models, the mixed model provided by the method has obvious superiority and is suitable for predicting sea wave height.
While specific embodiments of the invention have been described above, it will be appreciated by those skilled in the art that this is by way of example only, and the scope of the invention is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the principles and spirit of the invention, but such changes and modifications fall within the scope of the invention.
Claims (3)
1. A method of predicting sea wave height, the method comprising:
performing VMD decomposition on the original sea wave height data sequence;
decomposing the VMD to obtain a plurality of discrete subsequences IMF 1 ，IMF 2 ……IMF k Inputting the result into an AM-LSTM model to obtain a model prediction result PIMF of the discrete subsequences 1 ，PIMF 2 ……PIMF k ；
Inputting the original sea wave height data sequence into the AM-LSTM model to obtain a model prediction result PIMF of the original sea wave height data sequence;
model predictive result PIMF for the plurality of discrete subsequences 1 ，PIMF 2 ……PIMF k Reconstructing and calculating the PIMF (model predictive model) of the original sea wave height data sequence to obtain the sea wave heightFinal predicted value of (2)The reconstruction calculation obtains the final predicted value of the sea wave height>Comprising the following steps:
and evaluating the final predicted value.
2. A method of ocean wave height prediction according to claim 1 wherein the VMD decomposing the original ocean wave height data sequence comprises:
initializing parameters of the VMDSetting the maximum iteration number n;
for the saidIteration is performed according to the following formula:
wherein lambda is Lagrangian multiplier and alpha is quadratic penaltyFactor τ is noise tolerance, i is imaginary number, k is the number of the plurality of discrete subsequences, w is center frequency, u k ，w k The kth discrete sub-sequence and the center frequency of said kth discrete sub-sequence,fourier transforms of f (t), values obtained by a multiplication operator alternating direction algorithm;
the iteration is carried out until the conditionUp to the point where the precision converges on criterion epsilon>0；
Output the final u k ，w k 。
3. A method of predicting sea wave height as claimed in claim 1, wherein the final predicted value is estimated using mean absolute error, root mean square error and mean absolute percentage error.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110313922.8A CN113033094B (en) | 2021-03-24 | 2021-03-24 | Sea wave height prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110313922.8A CN113033094B (en) | 2021-03-24 | 2021-03-24 | Sea wave height prediction method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113033094A CN113033094A (en) | 2021-06-25 |
| CN113033094B true CN113033094B (en) | 2024-02-09 |
Family
ID=76473384
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110313922.8A Active CN113033094B (en) | 2021-03-24 | 2021-03-24 | Sea wave height prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113033094B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114519806B (en) * | 2022-01-29 | 2022-10-11 | 国家海洋环境预报中心 | Ocean wave level observation model training method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107885951A (en) * | 2017-11-27 | 2018-04-06 | 河海大学 | A kind of Time series hydrological forecasting method based on built-up pattern |
| CN109118000A (en) * | 2018-08-07 | 2019-01-01 | 广东工业大学 | A kind of short-term wind speed forecasting method based on CEEMD-VMD-GA-ORELM model |
| CN111461416A (en) * | 2020-03-23 | 2020-07-28 | 上海电气风电集团股份有限公司 | Wind speed prediction method, system, electronic device and storage medium |
| CN112307676A (en) * | 2020-11-04 | 2021-02-02 | 国家海洋局北海预报中心((国家海洋局青岛海洋预报台)(国家海洋局青岛海洋环境监测中心站)) | Wave height numerical prediction model result correction method |
-
2021
- 2021-03-24 CN CN202110313922.8A patent/CN113033094B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107885951A (en) * | 2017-11-27 | 2018-04-06 | 河海大学 | A kind of Time series hydrological forecasting method based on built-up pattern |
| CN109118000A (en) * | 2018-08-07 | 2019-01-01 | 广东工业大学 | A kind of short-term wind speed forecasting method based on CEEMD-VMD-GA-ORELM model |
| CN111461416A (en) * | 2020-03-23 | 2020-07-28 | 上海电气风电集团股份有限公司 | Wind speed prediction method, system, electronic device and storage medium |
| CN112307676A (en) * | 2020-11-04 | 2021-02-02 | 国家海洋局北海预报中心((国家海洋局青岛海洋预报台)(国家海洋局青岛海洋环境监测中心站)) | Wave height numerical prediction model result correction method |
Non-Patent Citations (2)
| Title |
|---|
| VMD小波能量的光伏逆变器软故障诊断;姜媛媛;《电力系统及其自动化学报》;论文第3-4页 * |
| 基于VMD 和双重注意力机制LSTM 的短期光伏功率预测;杨显晶等;《电力系统自动化》;论文第1-9页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113033094A (en) | 2021-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Smart wind speed forecasting approach using various boosting algorithms, big multi-step forecasting strategy | |
| Liang et al. | A novel wind speed prediction strategy based on Bi-LSTM, MOOFADA and transfer learning for centralized control centers | |
| Wang et al. | A review of deep learning for renewable energy forecasting | |
| Ma et al. | A hybrid attention-based deep learning approach for wind power prediction | |
| CN113392961B (en) | Method for extracting mesoscale eddy track stable sequence and predicting cyclic neural network | |
| CN110070229A (en) | The short term prediction method of home electrical load | |
| Zhou et al. | Short-term wind power prediction optimized by multi-objective dragonfly algorithm based on variational mode decomposition | |
| CN109242212A (en) | A kind of wind-powered electricity generation prediction technique based on change Mode Decomposition and length memory network | |
| Jun et al. | A new weighted CEEMDAN-based prediction model: An experimental investigation of decomposition and non-decomposition approaches | |
| CN107885951A (en) | A kind of Time series hydrological forecasting method based on built-up pattern | |
| CN111950759B (en) | Short-term wind speed prediction method based on two-stage decomposition, LSTM and AT | |
| CN111144644B (en) | Short-term wind speed prediction method based on variation variance Gaussian process regression | |
| CN113705864A (en) | Weather drought prediction method and device based on VMD-CNN-BilSTM-ATT mixed model | |
| Zhang et al. | Interval prediction of ultra-short-term photovoltaic power based on a hybrid model | |
| CN112434891A (en) | Method for predicting solar irradiance time sequence based on WCNN-ALSTM | |
| CN105760961A (en) | Photovoltaic output power prediction method and system | |
| Kaur et al. | Wind speed forecasting using hybrid wavelet transform-ARMA techniques | |
| Fu et al. | Multi-step-ahead significant wave height prediction using a hybrid model based on an innovative two-layer decomposition framework and LSTM | |
| CN116822374A (en) | Wind speed multi-step prediction method and device, computer equipment and storage medium | |
| CN113033094B (en) | Sea wave height prediction method | |
| CN109669169B (en) | Method for detecting weak target signal under sea clutter background | |
| Sun et al. | Multi-step wind speed forecasting based on a hybrid decomposition technique and an improved back-propagation neural network | |
| CN117407660B (en) | Regional sea wave forecasting method based on deep learning | |
| Shi et al. | Deep-learning-based wind speed forecasting considering spatial–temporal correlations with adjacent wind turbines | |
| Pang et al. | A novel multivariable hybrid model to improve short and long-term significant wave height prediction |
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 | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Lu Peng Inventor after: Nian Shengquan Inventor after: Cao Yang Inventor after: Zhang Na Inventor after: Liu Kaibin Inventor after: Wang Zhenhua Inventor after: Zheng Zongsheng Inventor before: Lu Peng Inventor before: Nian Shengquan Inventor before: Cao Yang Inventor before: Zhang Na Inventor before: Liu Kaibin Inventor before: Wang Zhenhua Inventor before: Liu Zongsheng |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant |