CN116449462B - Method, system, storage medium and equipment for predicting effective wave height space-time sequence of sea wave - Google Patents
Method, system, storage medium and equipment for predicting effective wave height space-time sequence of sea wave Download PDFInfo
- Publication number
- CN116449462B CN116449462B CN202310720313.3A CN202310720313A CN116449462B CN 116449462 B CN116449462 B CN 116449462B CN 202310720313 A CN202310720313 A CN 202310720313A CN 116449462 B CN116449462 B CN 116449462B
- Authority
- CN
- China
- Prior art keywords
- data
- wave
- moment
- wave height
- prediction
- 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 43
- 238000013528 artificial neural network Methods 0.000 claims abstract description 32
- 230000000306 recurrent effect Effects 0.000 claims abstract description 24
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 6
- 230000015654 memory Effects 0.000 claims description 6
- 230000002035 prolonged effect Effects 0.000 abstract description 3
- 238000012360 testing method Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 2
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008878 coupling Effects 0.000 description 2
- 238000010168 coupling process Methods 0.000 description 2
- 238000005859 coupling reaction Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 230000009897 systematic effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01W—METEOROLOGY
- G01W1/00—Meteorology
- G01W1/10—Devices for predicting weather conditions
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C13/00—Surveying specially adapted to open water, e.g. sea, lake, river or canal
-
- 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
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- 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
-
- 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)
- Life Sciences & Earth Sciences (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Environmental & Geological Engineering (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Atmospheric Sciences (AREA)
- Remote Sensing (AREA)
- Environmental Sciences (AREA)
- Radar, Positioning & Navigation (AREA)
- Biodiversity & Conservation Biology (AREA)
- Ecology (AREA)
- Hydrology & Water Resources (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of sea wave forecasting, and discloses a method, a system, a storage medium and equipment for forecasting effective wave height space-time sequence of sea waves, wherein the method comprises the following steps: acquiring effective wave height of sea waves from an initial moment to a current moment and multi-element data from a second moment to a first future moment; taking the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment as input data sequences, and predicting to obtain the effective wave height data of the sea wave from the first future moment by adopting a prediction recurrent neural network; and obtaining the wave effective wave height data from the second future time to the target future time through single-step prediction based on a recursion strategy and adopting a prediction recursion neural network based on the wave effective wave height data from the first future time and the multi-element data from the second future time to the target future time. The prediction error is stabilized, the prediction precision of the effective wave height of the sea wave is effectively improved, and the effective prediction time is prolonged.
Description
Technical Field
The invention relates to the technical field of sea wave forecasting, in particular to a method, a system, a storage medium and equipment for forecasting effective wave height space-time sequence of sea waves.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
Ocean waves are the most common phenomenon in the ocean, and ocean waves of extreme height are considered as ocean disasters threatening offshore operations and sailing, so ocean wave forecasting has become an indispensable work for maritime institutions worldwide. The wave effective wave height is used as one of the most important parameters to describe the statistical distribution characteristics of wave heights, and the traditional prediction of wave effective wave height mainly uses a numerical mode method based on a Navier-Stokes equation, for example, a third generation numerical wave model includes SWAN (Simulating Waves Nearshore, third generation coast wave numerical calculation mode) and wave watch III (third generation wave numerical prediction mode). These models use a large amount of computational resources for discrete calculations rather than differential equations, resulting in unavoidable systematic errors. In addition, the defects that the numerical mode is insufficient in integrating new observation system data and cannot effectively apply massive observation data are gradually revealed. In contrast, the development of big data and artificial intelligence technology in recent years provides a new data driven approach to sea wave prediction. In particular, deep learning is of increasing interest for use in sea wave prediction, however, most deep learning methods are designed for single point prediction, which inevitably reduces the accuracy of effective wave height prediction of sea waves at the target location, since the wave heights are a two-dimensional field, spatially cross-linked.
In order to solve the problem of space-time sequence prediction of the effective wave height of the sea wave, a space-time sequence prediction algorithm ConvLSTM (Convolutional LSTM Network, convolution LSTM network) is applied to the two-dimensional time sequence effective wave height prediction, but the current effective wave height space-time sequence prediction of the sea wave lacks the constraint of multiple power elements, the error increases rapidly along with the time, and the effective prediction time is limited to be within 24 hours.
Disclosure of Invention
In order to solve the problems, the invention provides a method, a system, a storage medium and equipment for predicting the effective wave height space-time sequence of sea waves, which use a multi-step prediction method based on a recursion strategy, and add future power elements and boundary data into prediction in the prediction process, so that the prediction error is stabilized, the effective wave height prediction precision of the sea waves is effectively improved, and the effective prediction time is prolonged.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
a first aspect of the present invention provides a method of wave effective wave height spatiotemporal sequence prediction comprising:
acquiring effective wave height of sea waves from an initial moment to a current moment and multi-element data from a second moment to a first future moment;
taking the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment as input data sequences, and predicting to obtain the effective wave height data of the sea wave from the first future moment by adopting a prediction recurrent neural network;
and obtaining the wave effective wave height data from the second future time to the target future time through single-step prediction based on a recursion strategy and adopting a prediction recursion neural network based on the wave effective wave height data from the first future time and the multi-element data from the second future time to the target future time.
Further, the multi-element data includes sea surface wind data, ocean current data, and water depth data.
Further, the single-step prediction based on the recursive strategy is specifically as follows: if the wave effective wave height data at the kth future time is predicted, adding the wave effective wave height data at the kth-1 future time and the multi-element data at the kth future time into the input data sequence, and after the wave effective wave height at the kth-1 time and the multi-element data at the kth time are moved out of the input data sequence, predicting by a prediction recurrent neural network to obtain the wave effective wave height data at the kth future time.
Further, after correcting the wave effective wave height data at the k-1 future time by adopting the boundary data of the wave effective wave height, adding the wave effective wave height data and the multi-element data at the k future time into the input data sequence.
Further, uniformly interpolating the acquired wave effective wave height from the initial moment to the current moment and the acquired multi-element data from the second moment to the first future moment by using a cubic spline interpolation method, and then taking the obtained multi-element data as an input data sequence.
Further, after normalizing the data in the input data sequence, predicting to obtain effective wave height data of the sea wave at the first future moment by adopting a prediction recurrent neural network.
Further, the accuracy of the prediction recurrent neural network to the effective wave height data prediction of the sea waves is measured by using the average absolute error, the root mean square error, the average absolute percentage error and the correlation coefficient.
A second aspect of the invention provides a wave effective wave height spatiotemporal sequence prediction system comprising:
a data acquisition module configured to: acquiring effective wave height of sea waves from an initial moment to a current moment and multi-element data from a second moment to a first future moment;
a first prediction module configured to: taking the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment as input data sequences, and predicting to obtain the effective wave height data of the sea wave from the first future moment by adopting a prediction recurrent neural network;
a second prediction module configured to: and obtaining the wave effective wave height data from the second future time to the target future time through single-step prediction based on a recursion strategy and adopting a prediction recursion neural network based on the wave effective wave height data from the first future time and the multi-element data from the second future time to the target future time.
A third aspect of the invention provides a computer readable storage medium having stored thereon a computer program for execution by a processor, the program when executed by the processor performing the steps in a method for wave effective wave height spatiotemporal sequence prediction as described above.
A fourth aspect of the invention provides a computer apparatus comprising a memory, a processor and a computer program stored on the memory and running on the processor, the processor implementing the steps in the wave effective wave height spatiotemporal sequence prediction method as described above when the program is executed.
Compared with the prior art, the invention has the beneficial effects that:
the invention provides a prediction method of effective wave height space-time sequence of sea waves, which uses a multi-step prediction method based on a recursion strategy, and adds future power elements and boundary data into prediction in the prediction process, so that the prediction error is stabilized, the effective wave height prediction precision of the sea waves is effectively improved, and the effective prediction time is prolonged.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention.
FIG. 1 is a block diagram of a PredRNN model according to a first embodiment of the present invention;
FIG. 2 is a two-dimensional scatter density chart of sea wave effective wave height and ERA-5 analysis data predicted by a sea wave effective wave height space-time sequence prediction method according to a first embodiment of the invention.
Detailed Description
The invention will be further described with reference to the drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the invention. 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.
The embodiments of the present invention and features of the embodiments may be combined with each other without conflict, and the present invention will be further described with reference to the drawings and embodiments.
Term interpretation:
LSTM: long Short-Term Memory network.
ST-LSTM: spatio-temporal LSTM.
PredRNN: predictive Recurrent Neural Network, a recurrent neural network is predicted.
Example 1
An object of the first embodiment is to provide a method for predicting the effective wave height space-time sequence of sea waves.
More and more researches have found that sea waves have strong coupling with ocean surface currents and the upper atmosphere, and meanwhile, close correlation exists between the effective wave height and the water depth of the sea waves. Thus, sea surface wind, current and water depth are very important for calculating land-based sea areas, such as the eastern Chinese sea area, and should be considered as inputs to the space-time ocean wave effective wave height prediction model.
According to the method for predicting the effective wave height space-time sequence of the sea wave, a multi-element-driven two-dimensional effective wave height prediction model with complete power process and perfect boundary is developed based on a neural network; using a multichannel PredRNN model (predicting recurrent neural network) and using a multi-step prediction strategy based on recursion to predict the effective wave height of the historical sea wave, 10m surface wind data, sea current data, water depth data and the effective wave height open boundary data of the sea wave for 1-72 hours continuously in the future; meanwhile, the accuracy of prediction of the effective wave height of the prediction recurrent neural network on the sea waves is measured by using an average absolute error (MAE), a Root Mean Square Error (RMSE), an average absolute percentage error (MAPE) and a Correlation Coefficient (CC).
The method for predicting the effective wave height space-time sequence of the sea wave comprises the following steps:
and step 1, acquiring the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment.
In this embodiment, the initial time is denoted as 1 time, the current time is denoted as t time, and the steps between the initial time and the current time include: second time (time 2), third time (time 3), …, time t-1 (time t-1).
In this embodiment, the several future times include: the first future time is denoted as t+1 time, the second future time is denoted as t+2 time, … …, the kth-1 future time is denoted as t+k-1 time, and the kth future time is denoted as t+k time (target future time); wherein k is greater than 1.
The effective wave height of the historical sea wave from time 1 to time t is expressed asThe method comprises the steps of carrying out a first treatment on the surface of the The multi-element data from time 2 to time t+1 is denoted +.>。
Wherein the multi-element data includes sea surface wind data (including U and V components), ocean current data (including U and V components), and water depth data; the sea surface wind data are 10m surface wind data.
And 2, taking the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment as an input data sequence, and predicting to obtain the effective wave height data of the sea wave from the first future moment by adopting a prediction recurrent neural network (PredRNN model).
As one embodiment, for the acquired multi-element data of the effective wave height of the sea wave from the initial time to the current time and the acquired multi-element data of the effective wave height of the sea wave from the second time to the first future time, a cubic spline interpolation method is used for uniformly interpolating the data to 1h×0.5 degrees of space-time resolution, and the data is normalized and then used as an input data sequence.
And step 3, obtaining the effective wave height data of the sea wave from the second future time to the target future time by single-step prediction based on a recursion strategy and adopting a prediction recursion neural network based on the effective wave height data of the sea wave from the first future time and the multi-element data from the second future time to the target future time.
The single-step prediction based on the recursion strategy is specifically as follows: and if the wave effective wave height data at the kth future moment is predicted, correcting the wave effective wave height data at the kth-1 future moment by adopting the boundary data of the wave effective wave height, adding an input data sequence together with the multi-element data at the kth future moment, and removing the wave effective wave height at the kth-1 moment and the multi-element data at the kth moment from the input data sequence, and predicting to obtain the wave effective wave height data at the kth future moment through a prediction recurrent neural network.
In the present embodiment, first, the effective wave height of the history sea wave from time 1 to time tAnd multi-element data from time 2 to time t+1->The combination is used as the input of a prediction recurrent neural network to predict the effective wave height data of the sea wave at the time t+1 +.>The method comprises the steps of carrying out a first treatment on the surface of the Then, the open boundary data +.>(boundary data of wave effective wave height) correction of wave effective wave height data at open boundary +.>The method comprises the steps of carrying out a first treatment on the surface of the Finally, will->And->Added to the input data sequence and +.>And->Shifting out the input data sequence to predict the effective wave height data of the sea wave at time t+2 +.>The method comprises the steps of carrying out a first treatment on the surface of the In this embodiment, it is necessary to predict the effective wave height data of the sea wave at 72 future times, that is, k=1, 2, …,72, so the above process loops and iteratively generates effective wave height prediction data of sea wave in the future for 1-72 hours.
In the present embodimentCorrection means using true open boundary dataFor predictive data->Replacement is performed, i.e. replacing the predicted value with the true value.
In this embodiment, the effective wave height of the ocean wave at the i-th moment and the multi-element data at the i+1-th moment form one element in the input data sequence, the input data sequence contains t elements, and the elements in the input data sequence are ordered according to time; where i=1, 2, …, t+k.
As shown in fig. 1, the PredRNN model includes t-layer network structures, each layer network structure includes 4 sequentially connected ST-LSTM, a q-th element in an input data sequence is input into the q-th layer network structure, and an output of the q-th layer network structure is input into the q+1-th layer network structure; the output of the jth ST-LSTM in each q-layer network structure is input to the jth ST-LSTM in the q+1-layer network structure. Where q=1, 2, …, t.
The boundary data of the effective wave height of the sea wave is the boundary data of the element of the effective wave height of the sea wave.
The multi-element data of the future moment comes from the prediction result of the numerical mode, and the prediction accuracy and the prediction effective duration of sea surface wind data, sea current data and water depth data are far higher than the effective wave height of sea waves.
In the embodiment, the effective wave height data of the sea wave is taken as a prediction object and an initial boundary, and meanwhile, the sea current data and the sea wind data which have strong coupling with the sea wave are taken as main power elements and boundaries to be added into the effective wave height prediction of the sea wave; and adding water depth data with strong correlation with the effective wave height of the sea wave in statistics as a lower boundary into effective wave height prediction of the sea wave.
In this embodiment, the prediction recurrent neural network has complete power elements (sea wind, ocean current) and complete boundary elements (water depth, ocean wave effective wave height open boundary data) in the training and prediction process; the multi-step prediction method based on the recursion strategy is used, future power factors and boundary data are added into the prediction in the prediction process, and the prediction error is stabilized.
In the embodiment, training and testing experiments of the predictive recurrent neural network are carried out by using data of 2011-2019 for ten years in time, and the range of the predictive recurrent neural network is 24-41 DEG N, 118-132 DEG E in space. Firstly, uniformly interpolating data into 1h multiplied by 0.5 DEG space-time resolution by using a cubic spline interpolation method, normalizing the data, wherein the processed data comprises a data matrix of 35 multiplied by 29 in space in time for 87,648 hours, and the data type comprises four data of effective wave height of sea waves, sea surface wind, sea currents and water depths, wherein stroke and sea current data respectively comprise data on two components of U and V, so the data dimension is (87648, 35, 29, 6); then, the data is divided into a training set (2011-2017), a verification set (2018) and a test set (2019) which are respectively used for predicting training verification and test of the recurrent neural network; finally, the data set is divided into data samples, formatted into an input-output format supported by the predictive recurrent neural network.
In the test experiments, the effective wave height of sea waves predicted based on the multielement-driven PredRNN model was within 0.2m in most areas with respect to the ERA-5 re-analyzed MAE in the first 12 hours, but increased over time; after 12 hours, MAE begins to accumulate and is mainly distributed at the eastern parts of Bohai sea and Korea peninsula, about 0.4m; after 24 hours the MAE was gradually stabilized with no significant difference between 24 and 72 hours. The accumulation of MAE in the eastern part of the Bohai sea and the Korean peninsula is likely due to the complexity of the land-sea environment and the land-surrounded sea area; on the other hand, MAE remained stable after 24 hours thanks to the complete dynamic process and boundaries of the model; wherein ERA-5 is the name of the dataset, which is known as the middle weather forecast center in Europe, and the dataset is analyzed, i.e. the true value in the experimental process.
Further, as shown in fig. 2, scatter plots and other evaluation indexes such as RMSE, MAPE, and CC are used to evaluate the accuracy of different prediction times, where the dashed line indicates that the predicted value is equal to the true value. For wave effective wave height prediction based on the multielement-driven PredRNN, CC was reduced from 0.99 at 1 hour to 0.95 at 12 hours, 0.90 at 24 hours and 0.87 at 72 hours, respectively, rmse was increased from 0.04m at 1 hour to 0.26m at 12 hours, 0.36m at 24 hours and 0.39m at 72 hours, respectively, which was superior to the existing condlstm algorithm-based wave effective wave height spatiotemporal sequence prediction.
Example two
An object of the second embodiment is to provide a wave height space-time sequence prediction system, including:
a data acquisition module configured to: acquiring effective wave height of sea waves from an initial moment to a current moment and multi-element data from a second moment to a first future moment;
a first prediction module configured to: taking the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment as input data sequences, and predicting to obtain the effective wave height data of the sea wave from the first future moment by adopting a prediction recurrent neural network;
a second prediction module configured to: and obtaining the wave effective wave height data from the second future time to the target future time through single-step prediction based on a recursion strategy and adopting a prediction recursion neural network based on the wave effective wave height data from the first future time and the multi-element data from the second future time to the target future time.
It should be noted that, each module in the embodiment corresponds to each step in the first embodiment one to one, and the implementation process is the same, which is not described here.
Example III
The present embodiment provides a computer-readable storage medium having stored thereon a computer program which when executed by a processor performs the steps in the wave effective wave height spatiotemporal sequence prediction method according to the above embodiment.
Example IV
The present embodiment provides a computer device comprising a memory, a processor and a computer program stored on the memory and running on the processor, the processor implementing the steps in the method for predicting wave height spatiotemporal sequence of ocean waves according to the above embodiment.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
While the foregoing description of the embodiments of the present invention has been presented in conjunction with the drawings, it should be understood that it is not intended to limit the scope of the invention, but rather, it is intended to cover all modifications or variations within the scope of the invention as defined by the claims of the present invention.
Claims (5)
1. The method for predicting the effective wave height space-time sequence of the sea wave is characterized by comprising the following steps of:
acquiring effective wave height of sea waves from an initial moment to a current moment and multi-element data from a second moment to a first future moment;
taking the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment as input data sequences, and predicting to obtain the effective wave height data of the sea wave from the first future moment by adopting a prediction recurrent neural network;
obtaining the wave effective wave height data from the second future time to the target future time through single-step prediction based on a recursion strategy and adopting a prediction recursion neural network based on the wave effective wave height data from the first future time and the multi-element data from the second future time to the target future time;
the multi-element data comprise sea surface wind data, ocean current data and water depth data;
the single-step prediction based on the recursion strategy is specifically as follows: if the wave effective wave height data at the kth future time is predicted, adding the wave effective wave height data at the kth-1 future time and the multi-element data at the kth future time into the input data sequence, and after the wave effective wave height at the kth-1 time and the multi-element data at the kth time are moved out of the input data sequence, predicting to obtain the wave effective wave height data at the kth future time through a prediction recurrent neural network;
correcting the wave effective wave height data at the k-1 future moment by adopting the boundary data of the wave effective wave height, and adding the wave effective wave height data and the multi-element data at the k future moment into the input data sequence;
uniformly interpolating the obtained wave effective wave height from the initial moment to the current moment and the obtained multi-element data from the second moment to the first future moment by using a cubic spline interpolation method, and then taking the obtained multi-element data as an input data sequence;
and after normalizing the data in the input data sequence, predicting by adopting a prediction recurrent neural network to obtain the effective wave height data of the sea wave at the first future moment.
2. The method for predicting the effective wave height space-time sequence of sea waves according to claim 1, wherein the accuracy of the prediction recurrent neural network for predicting the effective wave height data of sea waves is measured by using average absolute error, root mean square error, average absolute percentage error and correlation coefficient.
3. The wave effective wave height space-time sequence prediction system is realized by the wave effective wave height space-time sequence prediction method as claimed in claim 1, and is characterized by comprising the following steps:
a data acquisition module configured to: acquiring effective wave height of sea waves from an initial moment to a current moment and multi-element data from a second moment to a first future moment;
a first prediction module configured to: taking the effective wave height of the sea wave from the initial moment to the current moment and the multi-element data from the second moment to the first future moment as input data sequences, and predicting to obtain the effective wave height data of the sea wave from the first future moment by adopting a prediction recurrent neural network;
a second prediction module configured to: and obtaining the wave effective wave height data from the second future time to the target future time through single-step prediction based on a recursion strategy and adopting a prediction recursion neural network based on the wave effective wave height data from the first future time and the multi-element data from the second future time to the target future time.
4. A computer readable storage medium having stored thereon a computer program for execution by a processor, wherein the program when executed by the processor performs the steps of the wave effective wave height spatiotemporal sequence prediction method of any of claims 1-2.
5. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor performs the steps in the wave-efficient wave-height spatiotemporal sequence prediction method of any of claims 1-2 when the program is executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310720313.3A CN116449462B (en) | 2023-06-19 | 2023-06-19 | Method, system, storage medium and equipment for predicting effective wave height space-time sequence of sea wave |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310720313.3A CN116449462B (en) | 2023-06-19 | 2023-06-19 | Method, system, storage medium and equipment for predicting effective wave height space-time sequence of sea wave |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116449462A CN116449462A (en) | 2023-07-18 |
| CN116449462B true CN116449462B (en) | 2023-10-03 |
Family
ID=87122355
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310720313.3A Active CN116449462B (en) | 2023-06-19 | 2023-06-19 | Method, system, storage medium and equipment for predicting effective wave height space-time sequence of sea wave |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116449462B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116822253B (en) * | 2023-08-29 | 2023-12-08 | 山东省计算中心（国家超级计算济南中心） | Hybrid precision implementation method and system suitable for MANUM sea wave mode |
| CN117574099B (en) * | 2024-01-16 | 2024-04-05 | 广东海洋大学 | Time-space correction deck upward wave time prediction method and system |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101789808B1 (en) * | 2017-01-06 | 2017-10-25 | (주)전략해양 | Ship operation index forecasting apparatus and the method thereof |
| KR20200084386A (en) * | 2018-12-20 | 2020-07-13 | 한동대학교 산학협력단 | Method for measuring significant wave height using artificial neural network in radar type wave gauge system |
| CN112711915A (en) * | 2021-01-08 | 2021-04-27 | 自然资源部第一海洋研究所 | Sea wave effective wave height prediction method |
| CN114549925A (en) * | 2022-01-18 | 2022-05-27 | 大连理工大学 | Sea wave effective wave height time sequence prediction method based on deep learning |
| CN115169439A (en) * | 2022-06-16 | 2022-10-11 | 中国人民解放军国防科技大学 | Method and system for predicting effective wave height based on sequence-to-sequence network |
| CN115222163A (en) * | 2022-09-20 | 2022-10-21 | 中国海洋大学 | Multi-factor medium-long term real-time forecasting method and system for harbor basin inlet waves and application |
-
2023
- 2023-06-19 CN CN202310720313.3A patent/CN116449462B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101789808B1 (en) * | 2017-01-06 | 2017-10-25 | (주)전략해양 | Ship operation index forecasting apparatus and the method thereof |
| KR20200084386A (en) * | 2018-12-20 | 2020-07-13 | 한동대학교 산학협력단 | Method for measuring significant wave height using artificial neural network in radar type wave gauge system |
| CN112711915A (en) * | 2021-01-08 | 2021-04-27 | 自然资源部第一海洋研究所 | Sea wave effective wave height prediction method |
| CN114549925A (en) * | 2022-01-18 | 2022-05-27 | 大连理工大学 | Sea wave effective wave height time sequence prediction method based on deep learning |
| CN115169439A (en) * | 2022-06-16 | 2022-10-11 | 中国人民解放军国防科技大学 | Method and system for predicting effective wave height based on sequence-to-sequence network |
| CN115222163A (en) * | 2022-09-20 | 2022-10-21 | 中国海洋大学 | Multi-factor medium-long term real-time forecasting method and system for harbor basin inlet waves and application |
Non-Patent Citations (2)
| Title |
|---|
| 基于LSTM神经网络的畸形波预测;赵勇;苏丹;邹丽;王爱民;;华中科技大学学报(自然科学版)(第07期);全文 * |
| 支持向量回归的机器学习方法在海浪预测中的应用;王燕;钟建;张志远;;海洋预报(第03期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116449462A (en) | 2023-07-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116449462B (en) | Method, system, storage medium and equipment for predicting effective wave height space-time sequence of sea wave | |
| CN111199270B (en) | Regional wave height forecasting method and terminal based on deep learning | |
| CN112561148A (en) | Ship track prediction method and system based on one-dimensional convolutional neural network and LSTM | |
| CN112446419A (en) | Time-space neural network radar echo extrapolation forecasting method based on attention mechanism | |
| CN112862774B (en) | Accurate segmentation method for remote sensing image building | |
| CN111967679A (en) | Ionized layer total electron content forecasting method based on TCN model | |
| CN113255972B (en) | Short-term rainfall prediction method based on Attention mechanism | |
| CN114445634A (en) | Sea wave height prediction method and system based on deep learning model | |
| CN114549925A (en) | Sea wave effective wave height time sequence prediction method based on deep learning | |
| CN115545334B (en) | Land utilization type prediction method and device, electronic equipment and storage medium | |
| CN113240169A (en) | Short-term rainfall prediction method of GRU network based on multi-mode data and up-down sampling | |
| CN115965862A (en) | SAR ship target detection method based on mask network fusion image characteristics | |
| CN115951430A (en) | Geomagnetic index forecasting method based on LSTM | |
| CN114004152B (en) | Multi-wind-field wind speed space-time prediction method based on graph convolution and recurrent neural network | |
| CN117233869B (en) | Site short-term wind speed prediction method based on GRU-BiTCN | |
| CN116341391B (en) | Precipitation prediction method based on STPM-XGBoost model | |
| CN112949944A (en) | Underground water level intelligent prediction method and system based on space-time characteristics | |
| CN113642591B (en) | Multi-beam submarine sediment layer type estimation method and system based on transfer learning | |
| CN116187563A (en) | Sea surface temperature space-time intelligent prediction method based on fusion improvement variation modal decomposition | |
| CN113344290B (en) | Method for correcting sub-season rainfall weather forecast based on U-Net network | |
| CN115540832A (en) | Satellite altimetry submarine topography correction method and system based on VGGNet | |
| CN113221450B (en) | Space-time prediction method and system for sparse non-uniform time sequence data | |
| CN115100543A (en) | Self-supervision self-distillation element learning method for small sample remote sensing image scene classification | |
| CN115222024B (en) | Short-term photovoltaic power generation prediction method and system based on depth feature selection network | |
| CN115062526B (en) | Three-dimensional ionosphere electron concentration distribution model training method based on deep learning |
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 |