Ship swing analysis system and method based on deep learning

Abstract

The invention relates to the field of marine equipment, and provides a ship swing analysis system and method based on deep learning, and the system comprises a data obtaining module which is used for obtaining time sequence data, and the time sequence data comprises ship driving time sequence data, ship swing time sequence data and meteorological time sequence data; the data division module is used for dividing the time series data into historical time series data and to-be-predicted time series data and determining data labels corresponding to the historical time series data according to the to-be-predicted time series data, and the data labels comprise rolling angle labels and pitching angle labels; the model training module is used for training a preset model according to the historical time series data and the data label to obtain a rolling angle prediction model and a pitching angle prediction model which can be used for determining to-be-predicted time series data according to the historical time series data; and the swing analysis module is used for predicting the effective rolling angle and the effective rolling angle of the ship based on the rolling angle prediction model and the pitching angle prediction model.

Description

Technical Field

[0001] This application relates to the field of marine equipment, and particularly to a ship rolling analysis system and method based on deep learning. Background Art

[0002] A ship will undergo various rolling motions such as rolling, pitching, and heaving in waves. When the rolling amplitude exceeds a certain limit, it may cause seasickness among the crew on the ship, a decrease in work efficiency, and even endanger the lives of crew members in extreme cases. For cargo, violent rolling may cause the cargo to shift and be damaged, thus affecting the stability and safety of the ship. In addition, the ship's equipment may also be damaged due to excessive rolling, affecting its normal operation. Therefore, in order to improve the safety of ships, it is extremely important to predict the motion response of ships under different sea conditions. By predicting ship rolling, measures can be taken in advance, such as adjusting the ship's speed, course, or using anti-rolling devices, to avoid danger to the ship. How to analyze and predict ship rolling has become an urgent problem to be solved. Summary of the Invention

[0003] The main purpose of this application is to provide a ship rolling analysis system based on deep learning, aiming to accurately predict the rolling of ships.

[0004] In a first aspect, this application provides a ship rolling analysis system based on deep learning, and the system includes: A data acquisition module, configured to acquire time-series data, where the time-series data includes: ship travel time-series data, ship rolling time-series data, and meteorological time-series data; A data division module, configured to divide the time-series data into historical time-series data and to-be-predicted time-series data, and determine the data labels corresponding to the historical time-series data according to the to-be-predicted time-series data, where the data labels include: roll angle label, pitch angle label; A model training module, configured to train a preset model according to the historical time-series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine the to-be-predicted time-series data based on the historical time-series data; A rolling analysis module, configured to predict the effective roll angle and effective roll angle of the ship based on the roll angle prediction model and the pitch angle prediction model.

[0005] In some embodiments, in the process of dividing the time-series data into historical time-series data and to-be-predicted time-series data by the data division module and determining the data labels corresponding to the historical time-series data according to the to-be-predicted time-series data, it is configured to: Based on the sliding window algorithm, determine the time-series data of the first preset length as the historical time-series data, and use the time-series data of the second preset length after the historical time-series data as the time-series data to be predicted; Use the roll angle data in the time-series data to be predicted as the roll angle label corresponding to the historical time-series data, and use the pitch angle data in the time-series data to be predicted as the pitch angle label corresponding to the historical time-series data.

[0006] In some embodiments, when the data acquisition module is used to acquire time-series data, the time-series data includes: ship driving time-series data, ship swaying time-series data, and meteorological time-series data, and it is used for: Determine the ship driving time-series data according to the ship driving data sampled by the ship control system and the time stamp corresponding to the ship driving data, where the ship driving data includes: speed data, heading data, and heading difference data; Determine the ship swaying time-series data according to the ship swaying data sampled by the on-board sensor and the time stamp corresponding to the ship swaying data, where the ship swaying data includes: roll angle data, pitch angle data; Acquire meteorological data and the time stamp corresponding to the meteorological data, and determine the meteorological time-series data, where the meteorological data includes: wind field data, current field data, and wave data.

[0007] In some embodiments, when the data acquisition module is used to acquire the meteorological data released by the meteorological center and the time stamp corresponding to the meteorological data, and determine the meteorological time-series data, it is used for: Acquire the original meteorological data released by the meteorological center and the time stamp corresponding to the original meteorological data; Perform time interpolation and spatial interpolation on the original meteorological data according to the time stamp to obtain the target meteorological data and the time stamp corresponding to the target meteorological data; Determine the meteorological time-series data according to the target meteorological data and the time stamp corresponding to the target meteorological data.

[0008] In some embodiments, the system further includes: A data correction module, which is used to correct the ship swaying data according to the ship driving data and the meteorological data, and determine the ship swaying time-series data according to the corrected ship swaying data.

[0009] In some embodiments, when the data correction module is used to correct the ship swaying data according to the ship driving data and the meteorological data, and determine the ship swaying time-series data according to the corrected ship swaying data, it is used for: Calculate the motion response amplitude of the ship based on the following formula: ; Among them, H is the wave height data, is the ship's natural frequency, Indicates the frequency of encounters between ships and waves; The ship roll data is corrected according to the motion response amplitude.

[0010] In some embodiments, the data correction module, in the process of correcting the ship swing data according to the ship driving data and the meteorological data, and determining the ship swing time series data according to the corrected ship swing data, is further configured to: The encounter frequency is calculated according to the following formula: ; in, represents the circular frequency of the waves, V represents the speed of the ship, α represents the angle between the wave direction and the ship's heading, and L represents the length of the ship.

[0011] In some embodiments, the data correction module, in the process of correcting the ship rolling data according to the motion response amplitude, is further configured to: Correcting the roll angle tag according to a roll motion response amplitude operator and the motion response amplitude; The pitch angle label is corrected according to a pitch motion response magnitude operator and the motion response amplitude.

[0012] In some embodiments, the deep learning-based ship sway analysis system further includes: The model verification module is used to determine the accuracy performance index and the generalization performance index of the roll angle prediction model and the pitch angle prediction model according to the effective roll angle output by the roll angle prediction model and the effective pitch angle output by the pitch angle prediction model.

[0013] In a second aspect, the present application further provides a ship sway analysis method based on deep learning, the method comprising: Acquiring time series data, the time series data including: ship travel time series data, ship swing time series data, and weather time series data; Dividing the time series data into historical time series data and time series data to be predicted, and determining data labels corresponding to the historical time series data according to the time series data to be predicted, wherein the data labels include: a roll angle label and a pitch angle label; Training a preset model based on the historical time series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine the time series data to be predicted based on the historical time series data; Predict the effective roll angle and effective pitch angle of the ship based on the roll angle prediction model and the pitch angle prediction model.

[0014] This application provides a ship rolling analysis system and method based on deep learning. The application includes a data acquisition module for acquiring time-series data, where the time-series data includes: ship driving time-series data, ship rolling time-series data, and meteorological time-series data; a data division module for dividing the time-series data into historical time-series data and to-be-predicted time-series data, and determining the data labels corresponding to the historical time-series data according to the to-be-predicted time-series data, where the data labels include: roll angle label, pitch angle label; a model training module for training a preset model according to the historical time-series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine the to-be-predicted time-series data based on the historical time-series data; a rolling analysis module for predicting the effective roll angle and effective roll angle of the ship based on the roll angle prediction model and the pitch angle prediction model. By separately constructing a roll angle prediction model and a pitch angle prediction model, the predictability of the effective roll angle and effective pitch angle is improved, which is beneficial to making preparations for the rolling of the ship in advance and improving the safety of ship navigation. Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions of the embodiments of this application, the drawings required for the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are some embodiments of this application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 It is a schematic block diagram of a ship rolling analysis system based on deep learning provided by an embodiment of this application; Figure 2 It is a schematic flow chart of a ship rolling analysis method based on deep learning provided by an embodiment of this application; Figure 3 It is a schematic block diagram of the structure of a computer device related to an embodiment of this application. Detailed Embodiments

[0017] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are some embodiments of this application, rather than all embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

[0018] The flowcharts shown in the accompanying drawings are only illustrative examples, not necessarily including all contents and operations / steps, nor necessarily executed in the described order. For example, some operations / steps can be decomposed, combined or partially merged, so the actual execution order may change according to the actual situation.

[0019] The embodiments of the present application provide a ship roll analysis system and method based on deep learning.

[0020] The following will describe in detail some embodiments of the present application with reference to the accompanying drawings. Without conflict, the following embodiments and the features in the embodiments can be combined with each other.

[0021] Please refer to Figure 1 , Figure 1 which is a schematic block diagram of a ship roll analysis system based on deep learning provided by an embodiment of the present application. The ship roll analysis system based on deep learning can be used in a terminal or a server to accurately predict the roll of a ship. Among them, the terminal can be an electronic device such as a mobile phone, a tablet computer, a notebook computer, a desktop computer, a personal digital assistant, and a wearable device; the server can be an independent server, a server cluster, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, Content Delivery Network (CDN), and big data and artificial intelligence platforms.

[0022] As Figure 1 shown, the ship roll analysis system based on deep learning includes a data acquisition module 110, a data division module 120, a model training module 130, and a roll analysis module 140.

[0023] The data acquisition module 110 is used to acquire time series data, and the time series data includes: ship travel time series data, ship roll time series data, and meteorological time series data; The data division module 120 is used to divide the time series data into historical time series data and to-be-predicted time series data, and determine the data labels corresponding to the historical time series data according to the to-be-predicted time series data. The data labels include: roll angle label, pitch angle label; The model training module 130 is used to train a preset model according to the historical time series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine the to-be-predicted time series data according to the historical time series data; The roll analysis module 140 is used to predict the effective roll angle and the effective roll angle of the ship based on the roll angle prediction model and the pitch angle prediction model.

[0024] Exemplarily, in the related art, the ship roll is usually analyzed in six degrees of freedom, which are: surge, sway, heave, roll, pitch, and yaw. Among them, roll and pitch are usually the most concerned degrees of freedom.

[0025] It can be understood that roll refers to the rotational movement of the ship around the longitudinal axis. In waves, roll is the most likely movement to occur, and the rolling amplitude is usually the largest. Roll may cause cargo displacement, damage to the hull structure, and even affect the stability of the ship, significantly affecting the safety of the ship; pitch refers to the rotational movement of the ship around the transverse axis. Pitch will cause the ship to pitch back and forth, which may affect the navigation performance of the ship and the comfort of passengers. In waves, pitch may cause the ship's deck to be flooded, increasing the risk of water ingress. Therefore, the ship roll analysis system based on deep learning provided by the embodiments of the present application predicts the effective roll angle and effective pitch angle of the ship by training a roll angle prediction model and a pitch angle prediction model.

[0026] Exemplarily, the roll movement of the ship is the result of the combination of the ship's navigation and the environmental impact brought by meteorological factors. Therefore, this solution trains the roll angle prediction model and the pitch angle prediction model through ship travel time series data, ship roll time series data, and meteorological time series data.

[0027] Exemplarily, the roll movement of the ship is also related to the ship itself. Ships with different structures and shapes will produce different roll movements in the same environment with the same travel parameters and headings. Therefore, the roll angle prediction model and the pitch angle prediction model can be trained separately for ships with different signals, and the trained roll angle prediction model and pitch angle prediction model are associated with the corresponding ship models.

[0028] In some embodiments, when the data partitioning module is used to partition the time series data into historical time series data and to-be-predicted time series data, and determine the data labels corresponding to the historical time series data according to the to-be-predicted time series data, it is used for: Based on the sliding window algorithm, determine the time series data of the first preset length as the historical time series data, and use the time series data of the second preset length after the historical time series data as the to-be-predicted time series data; Use the roll angle data in the to-be-predicted time series data as the roll angle label corresponding to the historical time series data, and use the pitch angle data in the to-be-predicted time series data as the pitch angle label corresponding to the historical time series data.

[0029] Exemplarily, during the process of model training, a part of the time series data is divided into historical time series data, and the historical time series data is used as the input of the model, so that the model predicts the time series data to be predicted according to the input, outputs the prediction result of the model, and optimizes the model by comparing the prediction result with the actual time series data to be predicted, minimizing the loss function, thereby obtaining the final roll angle prediction model and pitch angle prediction model.

[0030] Exemplarily, the ship's swing time series data includes the roll angle data and pitch angle data of the ship. Since the main purpose of the embodiments of the present application is to train models that can predict the roll angle and pitch angle respectively, the roll angle data and pitch angle data are respectively used as the labels of the historical time series data.

[0031] In some embodiments, when the data acquisition module is used to acquire time series data, the time series data includes: ship travel time series data, ship swing time series data, and meteorological time series data, and it is used for: Determine the ship travel time series data according to the ship travel data sampled by the ship control system and the time stamp corresponding to the ship travel data, where the ship travel data includes: speed data, heading data, and heading difference data; Determine the ship swing time series data according to the ship swing data sampled by the on-board sensor and the time stamp corresponding to the ship swing data, where the ship swing data includes: roll angle data, pitch angle data; Obtain meteorological data and the time stamp corresponding to the meteorological data, and determine the meteorological time series data, where the meteorological data includes: wind field data, flow field data, and wave data.

[0032] Exemplarily, the GPS module in the ship control system can obtain the position information of the ship in real time, and calculate the speed of the ship relative to the ground by calculating the distance and time interval between consecutive positions as the speed data; the magnetic compass or gyrocompass of the ship can measure the actual heading of the ship in real time as the heading data; the automatic steering system of the ship will compare the actual heading measured by the compass with the set command heading, and calculate the difference between the actual heading and the command heading as the heading difference data.

[0033] Exemplarily, by setting gyroscopes at multiple positions on the ship, the data detected by the gyroscopes can be calculated to obtain the roll angle data and pitch angle data as the ship swing data. Exemplarily, in addition to the roll angle data and pitch angle data, the ship swing data may also include data of other degrees of freedom, which are not limited herein.

[0034] Exemplarily, meteorological data can be extracted from the database published by the meteorological center. For example, wind field data can be extracted from the fifth-generation global reanalysis dataset ERA5 provided by the European Centre for Medium-Range Weather Forecasts (ECMWF). ERA5 covers multiple variables of the global atmosphere, land, and ocean from 1940 to the present. It combines model data with global observation data and generates globally consistent climate data in space and time through data assimilation technology. Similarly, flow field data can be obtained from the Sub-Mesoscale Ocean Circulation (SMOC) data published by the meteorological center, which covers relevant data on sub-mesoscale circulation phenomena in a large number of oceans around the world.

[0035] Exemplarily, wave data can be determined by monitoring the sponge with the on-board equipment of the ship. For example, it can include wave height data, wave period data, etc.

[0036] In some embodiments, when the data acquisition module is used to acquire the meteorological data published by the meteorological center and the time stamp corresponding to the meteorological data to determine the meteorological time series data, it is used for: Acquire the original meteorological data published by the meteorological center and the time stamp corresponding to the original meteorological data; Perform time interpolation and spatial interpolation on the original meteorological data according to the time stamp to obtain the target meteorological data and the time stamp corresponding to the target meteorological data; Determine the meteorological time series data according to the target meteorological data and the time stamp corresponding to the target meteorological data.

[0037] Exemplarily, the original meteorological data of the meteorological center usually has a large scale and low resolution. For example, the spatial resolution of the ERA5 dataset at the sea surface is 0.25°×0.25° (in units of longitude and latitude). Such a resolution enables the dataset to cover the global ocean and provide a general description of the flow field of the global ocean. However, when used for ship roll analysis, since the size of the ship is relatively small compared to the scale of the resolution, there are likely to be inaccurate problems. Therefore, it is necessary to perform interpolation processing on the flow field data in the spatial scale. Specifically, according to the flow field data at two adjacent positions, determine the flow field data at the midpoint position between the two positions, thereby improving the resolution of the data.

[0038] Similarly, if the resolution of the original meteorological data in the time scale is low, interpolation processing can also be performed on the original meteorological data in the time scale, which will not be elaborated here.

[0039] Exemplarily, after interpolating the original meteorological data, the target meteorological data is obtained, and the basis for constructing the meteorological time series data is formed through the target meteorological data.

[0040] In some embodiments, the system further includes: A data correction module, configured to correct the ship sway data according to the ship travel data and meteorological data, and determine the ship sway time series data according to the corrected ship sway data.

[0041] Exemplarily, the ship sway data is obtained by fitting the sensor data monitored by sensors arranged at various positions on the ship. However, during the measurement process, the sensors may be interfered by various factors, such as environmental noise, the drift of the sensors themselves, the deviation of the installation position, etc., resulting in errors in the measurement results. Therefore, it is necessary to correct the ship sway data according to the ship travel data and meteorological data to make the ship sway data correspond to the current ship travel data and meteorological data.

[0042] Moreover, through the data correction method provided by the embodiments of the present application, the ship sway data measured by the sensors can also be made more corresponding to the meteorological data measured by the meteorological center.

[0043] In some embodiments, during the process of using the data correction module to correct the ship sway data according to the ship travel data and meteorological data, and determine the ship sway time series data, it is configured to: Calculate the motion response amplitude of the ship based on the following formula: ; where H is the wave height data, is the natural frequency of the ship, represents the encounter frequency between the ship and the sea waves; Correct the ship sway data according to the motion response amplitude.

[0044] Exemplarily, the motion response amplitude is an index to measure the severity of the ship's motion in waves, used to quantify the seakeeping performance of the ship under different sea conditions. Among them, the motion response amplitude is related to the sea waves, reflecting the wave height data and the encounter frequency of the sea waves in the formula. It is also related to the nature of the ship itself, reflected in the natural frequency of the ship in the formula.

[0045] Exemplarily, represents the inherent characteristic data of the ship, that is, the natural frequency of the ship, which refers to the frequency of the free vibration of the ship's structure or motion system when there is no external excitation (such as waves, wind force, etc.). It reflects the natural vibration characteristics of the ship's structure or motion system and can be calculated according to parameters such as the width of the ship, the displacement of the ship, the depth of the ship's hull, and the length between perpendiculars, or can also be determined according to actual measurements or data provided by the manufacturer.

[0046] In some embodiments, during the process of the data correction module for correcting the ship sway data according to the ship travel data and meteorological data and determining the ship sway time series data based on the corrected ship sway data, it is further configured to: Calculate the encounter frequency according to the following formula: ; Wherein, represents the circular frequency of the sea wave, V represents the ship's speed, α represents the angle between the sea wave direction and the ship's heading, and L represents the ship's length.

[0047] Exemplarily, the motion response amplitude is also related to the ship travel data, which is reflected in the ship's speed and the angle between the sea wave direction and the ship's heading in the formula; it is also related to the nature of the ship itself, which is reflected in the ship's length in the formula. Among them, the circular frequency of the sea wave can be determined by the flow field data, and the ship's speed, the angle between the sea wave direction and the ship's heading can be determined according to the ship travel data.

[0048] In some embodiments, during the process of the data correction module for correcting the ship sway data according to the motion response amplitude, it is further configured to: Correct the roll angle tag according to the roll motion response amplitude operator and the motion response amplitude; Correct the pitch angle tag according to the pitch motion response amplitude operator and the motion response amplitude.

[0049] Exemplarily, the relationship between the roll angle and the pitch angle and the motion response amplitude can be described by the ship's motion response amplitude operator (Response Amplitude Operator, RAO). RAO is the ratio of the motion response amplitude of a certain degree of freedom of the ship to the wave amplitude under the action of a regular wave with a unit wave amplitude, and is used to describe the motion response characteristics of the ship in waves.

[0050] Therefore, the theoretical roll angle can be calculated based on the roll RAO and the motion response amplitude. When the difference between the theoretical roll angle and the roll angle data is greater than the preset threshold, the roll angle data is corrected according to the theoretical roll angle. For example, the average value of the theoretical roll angle and the roll angle data is used as the roll angle tag.

[0051] Similarly, the pitch angle tag is corrected according to the pitch motion response amplitude operator and the motion response amplitude.

[0052] In some embodiments, the deep learning-based ship sway analysis system further includes: A model verification module is used to determine the accuracy performance index and generalization performance index of the roll angle prediction model and the pitch angle prediction model based on the effective roll angle output by the roll angle prediction model and the effective pitch angle output by the pitch angle prediction model.

[0053] Exemplarily, the roll angle prediction model and the pitch angle prediction model are verified on the validation data set, so as to determine the accuracy score and generalization score of the roll angle prediction model and the pitch angle prediction model, and obtain the accuracy performance index and generalization performance index.

[0054] Exemplarily, the roll angle prediction model and the pitch angle prediction model can be trained by the method of "Gradient Boosting Tree" (abbreviated as GBM), specifically, they can be trained by LightGBM.

[0055] A ship rolling analysis system based on deep learning provided by the present application includes a data acquisition module for acquiring time series data, where the time series data includes: ship driving time series data, ship rolling time series data, and meteorological time series data; a data division module for dividing the time series data into historical time series data and to-be-predicted time series data, and determining the data labels corresponding to the historical time series data according to the to-be-predicted time series data, where the data labels include: roll angle label, pitch angle label; a model training module for training a preset model according to the historical time series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model capable of determining the to-be-predicted time series data based on the historical time series data; a rolling analysis module for predicting the effective roll angle and the effective roll angle of the ship based on the roll angle prediction model and the pitch angle prediction model. By respectively constructing a roll angle prediction model and a pitch angle prediction model, the predictability of the effective roll angle and the effective pitch angle is improved, which is beneficial to making preparations for the rolling of the ship in advance and improving the safety of ship navigation.

[0056] Please refer to Figure 2 , Figure 2 which is a schematic flow chart of a ship rolling analysis method based on deep learning provided by an embodiment of the present application.

[0057] An embodiment of the present application further provides a ship rolling analysis method based on deep learning, and the method includes steps S101 - S104: Step S101, acquire time series data, where the time series data includes: ship driving time series data, ship rolling time series data, and meteorological time series data; Step S102: Divide the time series data into historical time series data and time series data to be predicted, and determine the data labels corresponding to the historical time series data according to the time series data to be predicted. The data labels include: roll angle label and pitch angle label. Step S103: Train a preset model according to the historical time series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine the time series data to be predicted according to the historical time series data. Step S104: Predict the effective roll angle and effective pitch angle of the ship based on the roll angle prediction model and the pitch angle prediction model.

[0058] In some embodiments, the dividing the time series data into historical time series data and time series data to be predicted, and determining the data labels corresponding to the historical time series data according to the time series data to be predicted includes: Based on the sliding window algorithm, determine the time series data of a first preset length as the historical time series data, and use the time series data of a second preset length after the historical time series data as the time series data to be predicted. Use the roll angle data in the time series data to be predicted as the roll angle label corresponding to the historical time series data, and use the pitch angle data in the time series data to be predicted as the pitch angle label corresponding to the historical time series data.

[0059] In some embodiments, the obtaining the time series data, where the time series data includes: ship driving time series data, ship swaying time series data, and meteorological time series data, includes: According to the ship driving data sampled by the ship control system and the time stamp corresponding to the ship driving data, determine the ship driving time series data, where the ship driving data includes: speed data, heading data, and heading difference data. According to the ship swaying data sampled by the on-board sensor and the time stamp corresponding to the ship swaying data, determine the ship swaying time series data, where the ship swaying data includes: roll angle data and pitch angle data. Obtain meteorological data and the time stamp corresponding to the meteorological data, and determine the meteorological time series data, where the meteorological data includes: wind field data, flow field data, and wave data.

[0060] In some embodiments, the obtaining the meteorological data released by the meteorological center and the time stamp corresponding to the meteorological data, and determining the meteorological time series data includes: Obtain the original meteorological data released by the meteorological center and the time stamp corresponding to the original meteorological data; Perform time interpolation and spatial interpolation on the original meteorological data according to the time stamp to obtain the target meteorological data and the time stamp corresponding to the target meteorological data; Determine the meteorological time series data according to the target meteorological data and the time stamp corresponding to the target meteorological data.

[0061] In some embodiments, the method further includes: Correct the ship roll data according to the ship travel data and the meteorological data, and determine the ship roll time series data according to the corrected ship roll data.

[0062] In some embodiments, the correcting the ship roll data according to the ship travel data and the meteorological data, and determining the ship roll time series data according to the corrected ship roll data includes: Calculate the motion response amplitude of the ship based on the following formula: ; where H is the wave height data, is the natural frequency of the ship, represents the encounter frequency of the ship and the sea wave; Correct the ship roll data according to the motion response amplitude.

[0063] In some embodiments, the correcting the ship roll data according to the ship travel data and the meteorological data, and determining the ship roll time series data according to the corrected ship roll data includes: Calculate the encounter frequency according to the following formula: ; where, represents the circular frequency of the sea wave, V represents the ship's speed, α represents the angle between the sea wave direction and the ship's heading, and L represents the ship's length.

[0064] In some embodiments, the correcting the ship roll data according to the motion response amplitude includes: Correct the roll angle label according to the roll motion response amplitude operator and the motion response amplitude; Correct the pitch angle label according to the pitch motion response amplitude operator and the motion response amplitude.

[0065] In some embodiments, the ship roll analysis method based on deep learning further includes: Determine the accuracy performance index and the generalization performance index of the roll angle prediction model and the pitch angle prediction model according to the effective roll angle output by the roll angle prediction model and the effective pitch angle output by the pitch angle prediction model.

[0066] Please refer to Figure 3 , Figure 3 , which is a schematic block diagram of the structure of a computer device provided by an embodiment of the present application. The computer device can be a server or a terminal.

[0067] As Figure 3 shown, the computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the memory can include a storage medium and an internal memory.

[0068] The storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any ship roll analysis method based on deep learning.

[0069] The processor is used to provide computing and control capabilities to support the operation of the entire computer device.

[0070] The internal memory provides an environment for the operation of the computer program in the storage medium. When the computer program is executed by the processor, the processor can execute any ship roll analysis system method based on deep learning.

[0071] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art can understand that Figure 3 the structure shown in

[0072] is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0073] Among them, in one embodiment, the processor is used to run the computer program stored in the memory to implement the following steps: Obtain time-series data, where the time-series data includes: ship travel time-series data, ship roll time-series data, and meteorological time-series data; Divide the time-series data into historical time-series data and to-be-predicted time-series data, and determine data labels corresponding to the historical time-series data according to the to-be-predicted time-series data. The data labels include: roll angle label, pitch angle label; Train a preset model according to the historical time-series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine the to-be-predicted time-series data based on the historical time-series data; Predict the effective roll angle and effective roll angle of the ship based on the roll angle prediction model and the pitch angle prediction model.

[0074] The embodiment of the present application also provides a computer-readable storage medium. A computer program is stored on the computer-readable storage medium. The computer program includes program instructions. The method implemented when the program instructions are executed can refer to various embodiments of the ship roll analysis method based on deep learning in the present application.

[0075] Among them, the computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk equipped on the computer device, a Smart Media Card (SMC), a Secure Digital (SD) card, a Flash Card, etc.

[0076] It should be understood that the terms used in the specification of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in the specification of the present application and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an", and "the" are intended to include the plural forms.

[0077] It should also be understood that the term "and / or" used in the description of the present application and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations. It should be noted that in this text, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or system comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such a process, method, article or system. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of additional identical elements in the process, method, article or system comprising such element.

[0078] The serial numbers of the embodiments of the present application above are only for description and do not represent the superiority or inferiority of the embodiments. The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A ship sway analysis system based on deep learning, characterized in that The system includes: A data acquisition module, configured to acquire time-series data, where the time-series data includes: ship travel time-series data, ship roll time-series data, and meteorological time-series data; A data division module, configured to divide the time-series data into historical time-series data and to-be-predicted time-series data, and determine data labels corresponding to the historical time-series data according to the to-be-predicted time-series data, where the data labels include: roll angle label, pitch angle label; A model training module, configured to train a preset model according to the historical time-series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine to-be-predicted time-series data according to historical time-series data; A roll analysis module, configured to predict the effective roll angle and the effective roll angle of the ship based on the roll angle prediction model and the pitch angle prediction model; A data correction module, configured to calculate the motion response amplitude of the ship based on the following formula: ； where H is the wave height data, is the natural frequency of the ship, represents the encounter frequency between the ship and the waves; And correct the ship roll data according to the motion response amplitude.

2. The ship sway analysis system based on deep learning according to claim 1, wherein In the process of dividing the time-series data into historical time-series data and to-be-predicted time-series data and determining data labels corresponding to the historical time-series data according to the to-be-predicted time-series data, the data division module is configured to: Based on a sliding window algorithm, determine time-series data of a first preset length as the historical time-series data, and use time-series data of a second preset length after the historical time-series data as the to-be-predicted time-series data; Use the roll angle data in the to-be-predicted time-series data as the roll angle label corresponding to the historical time-series data, and use the pitch angle data in the to-be-predicted time-series data as the pitch angle label corresponding to the historical time-series data.

3. The ship sway analysis system based on deep learning according to claim 1, characterized in that In the process of acquiring time-series data, where the time-series data includes: ship travel time-series data, ship roll time-series data, and meteorological time-series data, the data acquisition module is configured to: Determine the ship travel time-series data according to the ship travel data sampled by the ship control system and the time stamp corresponding to the ship travel data, where the ship travel data includes: speed data, heading data, heading difference data; Determine the ship roll time-series data according to the ship roll data sampled by the on-board sensor and the time stamp corresponding to the ship roll data, where the ship roll data includes: roll angle data, pitch angle data; Acquire meteorological data and the time stamp corresponding to the meteorological data, and determine the meteorological time-series data, where the meteorological data includes: wind field data, flow field data, wave data.

4. The ship sway analysis system based on deep learning according to claim 3, characterized in that In the process of acquiring meteorological data released by the meteorological center and the time stamp corresponding to the meteorological data and determining the meteorological time-series data, the data acquisition module is configured to: Acquire the original meteorological data released by the meteorological center and the time stamp corresponding to the original meteorological data; Perform time interpolation and spatial interpolation on the original meteorological data according to the time stamp to obtain target meteorological data and the time stamp corresponding to the target meteorological data; Determine the meteorological time-series data according to the target meteorological data and the time stamp corresponding to the target meteorological data.

5. The ship rolling analysis system based on deep learning according to claim 3, characterized in that, The data correction module is further configured to: Correct the ship's roll data based on the ship's travel data and meteorological data, and determine the ship's roll time series data based on the corrected ship's roll data.

6. The ship sway analysis system based on deep learning according to claim 5, characterized in that During the process of the data correction module for correcting the ship's roll data according to the ship's travel data and meteorological data, and determining the ship's roll time series data based on the corrected ship's roll data, it is also used for: Calculate the encounter frequency according to the following formula: ； wherein, represents the circular frequency of the sea wave, V represents the ship speed, α represents the included angle between the sea wave direction and the ship course, and L represents the ship length.

7. The ship sway analysis system based on deep learning according to claim 5, characterized in that, During the process of the data correction module for correcting the ship's roll data according to the motion response amplitude, it is also used for: Correct the roll angle label according to the roll motion response amplitude operator and the motion response amplitude; Correct the pitch angle label according to the pitch motion response amplitude operator and the motion response amplitude.

8. The ship rolling analysis system based on deep learning according to any one of claims 1-7, characterized in that, The deep learning-based ship roll analysis system further includes: A model verification module for determining the accuracy performance index and generalization performance index of the roll angle prediction model and the pitch angle prediction model based on the effective roll angle output by the roll angle prediction model and the effective pitch angle output by the pitch angle prediction model.

9. A ship sway analysis method based on deep learning, characterized in that, The method includes: Obtain time series data, where the time series data includes: ship travel time series data, ship roll time series data, and meteorological time series data; Divide the time series data into historical time series data and to-be-predicted time series data, and determine the data labels corresponding to the historical time series data according to the to-be-predicted time series data. The data labels include: roll angle label, pitch angle label; Train a preset model according to the historical time series data and the data labels to obtain a roll angle prediction model and a pitch angle prediction model that can be used to determine the to-be-predicted time series data based on the historical time series data; Predict the effective roll angle and effective roll angle of the ship based on the roll angle prediction model and the pitch angle prediction model; Calculate the motion response amplitude of the ship according to the following formula: ； where H is the wave height data, is the natural frequency of the ship, represents the encounter frequency between the ship and the ocean waves; and correct the ship rolling data according to the amplitude of the motion response.

Images