CN115470957A

CN115470957A - System, computer equipment and storage medium for predicting wave height of offshore waves during typhoon based on deep learning

Info

Publication number: CN115470957A
Application number: CN202110647400.1A
Authority: CN
Inventors: 孟凡; 宋弢; 谢鹏飞; 李颖; 庞皓月; 王家荣
Original assignee: China University of Petroleum East China
Current assignee: China University of Petroleum East China
Priority date: 2021-06-10
Filing date: 2021-06-10
Publication date: 2022-12-13

Abstract

The invention discloses a deep learning-based system for predicting wave height of offshore waves during a typhoon, and belongs to the technical field of space-time sequence prediction. The system comprises a bidirectional gated cyclic unit (BiGRU) model, wherein the bidirectional gated cyclic unit model comprises a sequence processing model consisting of two gated cyclic units (GRUs), one input is forward input, the other input is reverse input, and the two gated cyclic unit model is a bidirectional recurrent neural network with only an input gate and a forgetting gate. The input of the model is real-time data acquired by a coastal buoy during typhoon and real-time data of typhoon, 6 parameters including real-time wave height, air pressure and wind speed acquired by an offshore buoy, lowest air pressure and real-time wind speed of a typhoon center, calculated distance between the typhoon center and the buoy and the like are used as model input, and the final output is predicted wave height in the future of 3 hours, 6 hours, 12 hours and 24 hours.

Description

System, computer equipment and storage medium for predicting wave height of offshore waves during typhoon based on deep learning

Technical Field

The invention relates to the technical field of space-time sequence prediction, in particular to a system, a computer device and a storage medium for predicting wave height of offshore waves during typhoon based on deep learning.

Background

Accurate wave height prediction plays an important role in ports, marine transportation, fisheries, various types of military operations, and the like, and the marine oil and gas industry has an important need for surrounding wave characteristics for a period of time in order to plan and perform safe and effective operations. Typhoon is one of the most destructive natural disasters in the world, and strong wind in a typhoon crossing environment carries a large amount of rainfall and huge storms, so that huge wave height is caused. Therefore, during a typhoon, the wave height prediction plays a very important role, and during the typhoon, the real-time prediction can be used for disaster preparation and early evacuation of various offshore operations, but during the typhoon, the wave height prediction for the typhoon also depends on a numerical mode method, which consumes a large amount of computing resources and has a long computing time, and the role of the wave height prediction is limited to an extreme weather event such as typhoon which the real-time performance is extremely high.

The oceans and the meteorological field gradually begin to carry out various prediction tasks by means of artificial intelligence technology, and a plurality of successful cases exist, in the aspect of long-term prediction of the early-nino phenomenon, the artificial intelligence technology exceeds all traditional methods, wave height in the typhoon period has extremely high requirements on instantaneity and accuracy, and the artificial intelligence recognition means is urgently needed to carry out rapid and effective prediction, so that a large amount of computing resources are not needed, and the extremely high instantaneity is realized.

Disclosure of Invention

Embodiments of the present invention provide a deep learning based wave height prediction system, computer device, storage medium for offshore waves during typhoons, and the following presents a brief summary in order to provide a basic understanding of some aspects of the disclosed embodiments. This summary is not an extensive overview and is intended to neither identify key/critical elements nor delineate the scope of such embodiments. Its sole purpose is to present some concepts in a simplified form as a prelude to the more detailed description that is presented later.

According to a first aspect of embodiments of the present invention, a deep learning based wave height prediction system for offshore waves during typhoons is provided.

In some alternative embodiments, the system includes a bidirectional gated cyclic unit (BiGRU) model that includes a sequential processing model of two gated cyclic units (GRUs), one input being a forward input and the other input being an inverse input, a two-way recurrent neural network with only input gates and forgetting gates. The input of the model is real-time data acquired by a coastal buoy during typhoon and real-time data of typhoon, 6 parameters including real-time wave height, air pressure and wind speed acquired by an offshore buoy, lowest air pressure and real-time wind speed of a typhoon center, calculated distance between the typhoon center and the buoy and the like are used as model input, and the final output is predicted wave height in the future of 3 hours, 6 hours, 12 hours and 24 hours.

Optionally, the gated loop unit (GRU) performs sufficient feature extraction on the multivariate time sequence, and continuously learns the long-term dependence relationship of the multivariate time sequence, which specifically includes: firstly, two gating states are obtained through the last transmitted state and the input of the current node, namely a reset gate (reset gate) for controlling reset and a update gate (update gate), after a gating signal is obtained, data after reset is obtained through the reset gate, the data and the input of the current node are spliced, the data are shrunk to the range of-1 to 1 through a hyperbolic tangent function, finally, the states are updated to the range of 0 to 1 through the update gate, and the more the gating signal is close to 1, the more data which represent memory are obtained.

Optionally, the bidirectional gated cyclic unit (BiGRU) model enables data to be input from both the positive direction and the negative direction, so that information at each time includes sequence information of previous and next times, that is, sequence information of a network at a specific time is increased, and information of historical data is fully utilized, thereby making prediction more accurate. The basic idea of BiGRU is to present each training sequence forward and backward to two separate hidden layers, both connected to the same output layer. The output layer thus has the complete past and future information for each point in the input sequence, the forward computation of BiGRU is the same as for unidirectional GRU, but the input sequence for bidirectional GRU is opposite for both hidden layers, and the output layer is not updated until both hidden layers have processed all the input sequence. The backward calculation of BiGRU is also similar to GRU, with all output layer entries calculated first and then returned to the hidden layers in two different directions.

Optionally, the input of the system is selected 6 variables, where the input variables include typhoon basic data issued by a meteorological department in real time, including minimum air pressure and real-time wind speed of a typhoon center, and position data (calculated distance between the typhoon center and a buoy); the input variables contain the instantaneous data acquired by the buoy including real-time wave height, air pressure and wind speed.

Optionally, the two-way gating circulation unit model is trained by using a large amount of existing typhoon observation data and a large amount of relevant data of the Chinese offshore buoy during typhoon, and perfect model parameters are obtained.

According to a second aspect of an implementation of the present invention, a computer device is provided.

In some optional embodiments, the computer device includes a memory, a graphics card, a central processing unit, and an executable program stored in the memory and capable of being processed by the central processing unit and the graphics card in parallel, wherein the central processing unit executes the program to implement the following steps: constructing a target detection and target segmentation model, wherein the target detection and target segmentation model comprises the following steps: the method comprises the steps that a characteristic extraction network and two branch networks comprise a target detection network and a target segmentation network, and firstly, the characteristic extraction network is utilized to carry out operation characteristic extraction such as convolution pooling on an input abdominal CT image; performing regression and classification on the extracted feature map by using a target detection network, and outputting the position and probability of gallstones; meanwhile, the image segmentation network performs up-sampling on the convolved and pooled feature images by deconvolution, restores the images to the size of the original image, finally obtains segmented contours of gallstones, namely masks, and can clearly and quantitatively analyze the features of the gallstones such as size, shape and the like.

Optionally, the gating cycle unit performs sufficient feature extraction on the multivariate time sequence, and continuously learns the long-term dependence relationship of the multivariate time sequence, which specifically includes: firstly, two gating states are obtained through the last transmitted state and the input of the current node, namely gating for controlling reset and gating for controlling update respectively, after a gating signal is obtained, data after reset is obtained through resetting gating, the data is spliced with the input of the current node, the data is shrunk to the range of-1 to 1 through a hyperbolic tangent function, finally, the states are updated to the range of 0 to 1 through the functions of forgetting and memorizing by using the updating gating, and the more the gating signal is close to 1, the more data representing the memory are.

The bidirectional gating cycle unit (BiGRU) model can enable data to be input from the positive direction and the negative direction simultaneously, information of each moment comprises sequence information of front and back moments, namely the sequence information of a network at a certain moment is increased, and information of historical data is fully utilized, so that prediction is more accurate. The basic idea of BiGRU is to present each training sequence forward and backward to two separate hidden layers, both connected to the same output layer. The output layer thus has the complete past and future information for each point in the input sequence, the forward computation of BiGRU is the same as for unidirectional GRU, but the input sequence for bidirectional GRU is opposite for both hidden layers, and the output layer is not updated until both hidden layers have processed all the input sequence. The backward calculation of BiGRU is also similar to GRU, with all output layer entries calculated first and then returned to the hidden layers in two different directions.

The technical scheme provided by the embodiment of the invention has the following beneficial effects:

the marine atmospheric field time-space sequence is intelligently processed by the aid of an intelligent technology, the difficult problems that a large amount of computing resources are consumed in a global numerical mode and real-time prediction cannot be achieved can be solved, wave height near a buoy can be rapidly and effectively determined, early warning of disaster evacuation and various marine operations can be assisted, processing speed is high, computing resources are small, and integration and large-scale application are facilitated.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.

FIG. 1 is a block diagram illustrating a deep learning based system for predicting wave height of offshore waves during a typhoon, according to an exemplary embodiment.

Detailed Description

The following description and the drawings sufficiently illustrate specific embodiments of the invention to enable those skilled in the art to practice them. Other embodiments may incorporate structural, logical, electrical, process, and other changes. The examples merely typify possible variations. Individual components and functions are optional unless explicitly required, and the sequence of operations may vary. Portions and features of some embodiments may be included in or substituted for those of others. The scope of embodiments of the invention encompasses the full ambit of the claims, as well as all available equivalents of the claims. Embodiments may be referred to herein, individually or collectively, by the term "invention" merely for convenience and without intending to voluntarily limit the scope of this application to any single invention or inventive concept if more than one is in fact disclosed. Herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, or apparatus. The use of the phrase "including a" does not exclude the presence of other, identical elements in a process, method or device that includes the recited elements, unless expressly stated otherwise. The embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. As for the methods, products and the like disclosed by the embodiments, the description is simple because the methods correspond to the method parts disclosed by the embodiments, and the related parts can be referred to the method parts for description.

Fig. 1 shows an alternative implementation architecture of an offshore wave height prediction system during a typhoon based on deep learning.

In this optional example, the system includes a bidirectional gated loop unit model, where the bidirectional gated loop unit model includes a sequence processing model composed of two gated loop units, one input is a forward input, the other input is a reverse input, and the input is a bidirectional recurrent neural network (bidirectional recurrent neural network) with only an input gate and a forgetting gate. The input of the model is real-time data acquired by a coastal buoy during typhoon and real-time data of typhoon, 6 parameters including real-time wave height, air pressure and wind speed acquired by an offshore buoy, lowest air pressure and real-time wind speed of a typhoon center, calculated distance between the typhoon center and the buoy and the like are used as model input, and the final output is predicted wave height in the future of 3 hours, 6 hours, 12 hours and 24 hours.

Optionally, the gating cycle unit performs sufficient feature extraction on the multivariate time series (multivariate time series), and continuously learns the long-term dependency relationship of the multivariate time series, which specifically includes: firstly, two gating states are obtained through the last transmitted state and the input of the current node, namely the gating for controlling reset and the gating for controlling update respectively, after a gating signal is obtained, data after reset is obtained through the reset gating, the data and the input of the current node are spliced, the data are shrunk to the range of-1 through a hyperbolic tangent function, finally, the states are updated to the range of 0-1 through the functions of forgetting and memorizing through the update gating, and the more the gating signal is close to 1, the more data which represent the memory are.

Optionally, the bidirectional gating cycle unit model enables data to be input from both positive and negative directions, so that information at each moment includes sequence information of previous and subsequent moments, which is equivalent to increase of sequence information of a network at a certain moment, and information of historical data is fully utilized, thereby enabling prediction to be more accurate. The basic idea of BiGRU is to present each training sequence forward and backward to two separate hidden layers, both connected to the same output layer. The output layer thus has complete past and future information for each point in the input sequence, the forward computation of BiGRU is the same as for unidirectional GRU, but the input sequence for bidirectional GRU is in the opposite direction for the two hidden layers, and the output layer is not updated until the two hidden layers have processed all the input sequences. The backward calculation of BiGRU is also similar to GRU, with all output layer entries calculated first and then returned to the hidden layers in two different directions.

Optionally, the bidirectional gated cyclic unit model is trained by using a large amount of existing typhoon observation data and a large amount of relevant data of the chinese offshore buoy during typhoon, and obtains perfect model parameters.

Optionally, the model further includes a training process of the bidirectional gated loop unit model, and a specific embodiment of the training process of the bidirectional gated loop unit model is provided below.

In the embodiment, in the training process of the target detection and target segmentation model, firstly, the real-time data of typhoon and the data acquired by the buoy are fused, the typhoon data are interpolated according to the time resolution of the buoy to form training data, the input layer is re-centered and re-scaled by Batch normalization (Batch normalization) in the training process to be normalized, so that the artificial neural network is faster and more stable, the number of cells of the BiGRU is set to be 64 (cell), then training is performed, the initial learning rate is 0.01, a loss function (loss function) is set to be mean absolute error loss (MAE loss), in the training process, errors between a predicted value and a real value are calculated, a network parameter is adjusted by an RMSprop optimizer, the weight of the model parameter is adjusted, and then loss function values are reduced through continuous iteration, so that the network is finally converged.

In an exemplary embodiment, a non-transitory computer readable storage medium, such as a memory, including instructions executable by a processor to perform steps of constructing a bi-directional gated loop unit (BiGRU) model including a sequential processing model of two gated loop units (GRUs), one input being a forward input and the other input being an inverse input, being an input-gate-only and forgetting-gate bi-directional recurrent neural network is also provided. The input of the model is real-time data acquired by a coastal buoy during typhoon and real-time data of typhoon, 6 parameters including real-time wave height, air pressure and wind speed acquired by an offshore buoy, lowest air pressure and real-time wind speed of a typhoon center, calculated distance between the typhoon center and the buoy and the like are used as model input, and the final output is predicted wave height in the future of 3 hours, 6 hours, 12 hours and 24 hours.

Optionally, the GRU performs sufficient feature extraction on the multivariate time sequence, and continuously learns the long-term dependency relationship of the multivariate time sequence, which specifically includes: firstly, two gating states are obtained through the last transmitted state and the input of the current node, namely gating for controlling reset and gating for controlling update respectively, after a gating signal is obtained, data after reset is obtained through resetting gating, the data is spliced with the input of the current node, the data is shrunk to the range of-1 to 1 through a hyperbolic tangent function, finally, the states are updated to the range of 0 to 1 through the functions of forgetting and memorizing by using the updating gating, and the more the gating signal is close to 1, the more data representing the memory are.

Optionally, the BiGRU model enables data to be input from both the positive and negative directions, so that information at each time includes sequence information of previous and subsequent times, which is equivalent to that the sequence information of the network at a certain specific time is increased, and information of historical data is fully utilized, thereby enabling prediction to be more accurate. The basic idea of BiGRU is to present each training sequence forward and backward to two separate hidden layers, both connected to the same output layer. The output layer thus has the complete past and future information for each point in the input sequence, the forward computation of BiGRU is the same as for unidirectional GRU, but the input sequence for bidirectional GRU is opposite for both hidden layers, and the output layer is not updated until both hidden layers have processed all the input sequence. The backward calculation of BiGRU is also similar to GRU, with all output layer entries calculated first and then returned to the hidden layers in two different directions.

The non-transitory computer readable storage medium may be a Read Only Memory (ROM), a Random Access Memory (RAMD), a magnetic tape, an optical storage device, and the like.

According to the invention, the time-space sequence in the ocean atmosphere field is intelligently processed by using an industrial intelligence technology, the difficult problems that a large amount of computing resources are consumed in an earth numerical mode and real-time prediction cannot be realized can be avoided, the wave height near the buoy can be quickly and effectively determined, early warning of disaster evacuation and various offshore operations can be assisted, the processing speed is high, the computing resources are small, and integration and large-scale application are facilitated.

Those of skill in the art would appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention. It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the embodiments disclosed herein, it should be understood that the disclosed methods, articles of manufacture (including but not limited to devices, apparatuses, etc.) may be implemented in other ways. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form. The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one position, or may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment. In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.

It should be understood that the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions. The present invention is not limited to the procedures and structures described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims

1. The system for predicting wave height of offshore waves during typhoon based on deep learning is characterized by comprising a bidirectional gated cyclic unit (BiGRU) model, wherein the bidirectional gated cyclic unit model comprises a sequence processing model consisting of two gated cyclic units (GRUs), one input is a forward input, the other input is a reverse input, and the two gated cyclic unit model is a bidirectional recurrent neural network with only an input gate and a forgetting gate. The input of the model is real-time data acquired by a coastal buoy during typhoon and real-time data of typhoon, 6 parameters including real-time wave height, air pressure and wind speed acquired by an offshore buoy, lowest air pressure and real-time wind speed of a typhoon center, calculated distance between the typhoon center and the buoy and the like are used as model input, and the final output is predicted wave height in the future of 3 hours, 6 hours, 12 hours and 24 hours.

2. The system according to claim 1, wherein the gated round robin unit (GRU) performs sufficient feature extraction on the multivariate time series to continuously learn long term dependencies of the multivariate time series, which specifically includes: firstly, two gating states are obtained through the last transmitted state and the input of the current node, namely a gate (resetgate) for controlling reset and a gate (update gate) for controlling update, after a gating signal is obtained, data after reset is obtained through the reset gating, the data and the input of the current node are spliced, the data are shrunk to the range of-1 through a hyperbolic tangent function, finally, the states are updated to the range of 0-1 through the update gating function, and the more gating signals are close to 1, the more data are represented to be memorized.

3. The system of claim 1, wherein the bidirectional gated round robin (BiGRU) model enables data to be input simultaneously from both the forward and reverse directions, such that the information at each time includes sequence information of preceding and following times, which is equivalent to the increase of sequence information of the network at a specific time, thereby making full use of the information of historical data and making prediction more accurate. The basic idea of BiGRU is to present each training sequence forward and backward to two separate hidden layers, both connected to the same output layer. The output layer thus has the complete past and future information for each point in the input sequence, the forward computation of BiGRU is the same as for unidirectional GRU, but the input sequence for bidirectional GRU is opposite for both hidden layers, and the output layer is not updated until both hidden layers have processed all the input sequence. The backward calculation of BiGRU is also similar to GRU, with all output layer entries calculated first and then returned to the hidden layers in two different directions.

4. The system of claim 1, wherein the inputs to the system are selected 6 variables, the input variables containing real-time typhoon-based data published by the meteorological department including the lowest pressure at the typhoon center and real-time wind speed, position data (calculated distance between the typhoon center and buoy); the input variables contain the instantaneous data acquired by the buoy including real-time wave height, air pressure and wind speed.

5. The system of claim 1, wherein the bi-directional gated cyclic unit model is trained using existing large typhoon observation data and large number of data associated with chinese offshore buoys during typhoons, and refined model parameters are obtained.

6. A computer device comprising a memory, a graphics card, a central processing unit, and an executable program stored in the memory and capable of being processed in parallel by the central processing unit and the graphics card, wherein the central processing unit implements the following steps when executing the program: and constructing a bidirectional gated cyclic unit (BiGRU) model, wherein the bidirectional gated cyclic unit model comprises a sequence processing model consisting of two gated cyclic units (GRUs), one input is forward input, the other input is reverse input, and the two-way recurrent neural network is a bidirectional recurrent neural network with only an input gate and a forgetting gate. The input of the model is real-time data acquired by a coastal buoy during typhoon and real-time data of typhoon, 6 parameters including real-time wave height, air pressure and wind speed acquired by an offshore buoy, lowest air pressure and real-time wind speed of a typhoon center, calculated distance between the typhoon center and the buoy and the like are used as model input, and the final output is predicted wave height in the future of 3 hours, 6 hours, 12 hours and 24 hours.

7. The computer apparatus of claim 7, wherein the gating cycle unit performs sufficient feature extraction on the multivariate time series to continuously learn long-term dependencies of the multivariate time series, and specifically comprises: firstly, two gating states are obtained through the last transmitted state and the input of the current node, namely gating for controlling reset and gating for controlling update respectively, after a gating signal is obtained, data after reset is obtained through resetting gating, the data is spliced with the input of the current node, the data is shrunk to the range of-1 to 1 through a hyperbolic tangent function, finally, the states are updated to the range of 0 to 1 through the functions of forgetting and memorizing by using the updating gating, and the more the gating signal is close to 1, the more data representing the memory are.

8. The computer device of claim 7, wherein the bidirectional gated cyclic unit model enables data to be input simultaneously from both the front and back directions, such that the information at each time includes sequence information of preceding and following times, which is equivalent to the increase of sequence information of the network at a specific time, and the information of historical data is fully utilized, thereby making prediction more accurate. The basic idea of the bi-directional gated loop cell model is to present each training sequence forward and backward to two separate hidden layers, both connected to the same output layer. The output layer has the complete past and future information for each point in the input sequence so that no replacement of the relevant target input occurs.

9. The computer apparatus of claim 7, wherein the inputs to the system are selected 6 variables, the input variables containing typhoon base data published in real time by the meteorological department, including the lowest air pressure and real-time wind speed at the typhoon center, position data (calculated distance between typhoon center and buoy); the input variables contain the instantaneous data acquired by the buoy including real-time wave height, air pressure and wind speed.

10. A storage medium storing a computer program, the computer program when executed by a central processing unit implementing the steps of: and constructing a bidirectional gated cyclic unit (BiGRU) model, wherein the bidirectional gated cyclic unit model comprises a sequence processing model consisting of two gated cyclic units (GRUs), one input is forward input, the other input is reverse input, and the two-way recurrent neural network is a bidirectional recurrent neural network with only an input gate and a forgetting gate. The input of the model is real-time data acquired by a coastal buoy during typhoon and real-time data of typhoon, 6 parameters including real-time wave height, air pressure and wind speed acquired by an offshore buoy, lowest air pressure and real-time wind speed of a typhoon center, calculated distance between the typhoon center and the buoy and the like are used as model input, and the final output is predicted wave height in the future of 3 hours, 6 hours, 12 hours and 24 hours.