Second-generation complete stability forecasting system and method for real sea navigation of ship

Abstract

The invention belongs to the field of ship navigation safety evaluation, and discloses a second-generation complete stability forecasting system and method for real sea navigation of a ship. The method comprises the following steps: acquiring actually measured marine environment parameters for sea wave forecasting, and acquiring ship navigation parameters for second-generation complete stability forecasting of a ship; sea wave forecasting is carried out, wherein regional meaningful wave height and average zero-crossing period forecasting of sea wave data is carried out based on an HP-VIT model; a pure stability loss sensitivity index forecasting model is constructed, and pure stability loss sensitivity index forecasting is completed; and a parameter rolling sensitive index forecasting model is constructed, and regional parameter rolling forecasting is completed. According to the method, rapid forecasting of the second-generation stability of the ship in the real sea environment can be achieved, the research result can provide auxiliary decision-making information for sailing safety guarantee, route planning correction and the like of the container ship, and the method has great significance in guaranteeing the safety of the ship and goods.

Description

Technical Field

[0001] The present invention belongs to the field of ship navigation safety assessment, and in particular relates to a second-generation intact stability prediction system and method for ships sailing in real seas. Background Art

[0002] With the development of the Maritime Silk Road, container ship traffic has grown rapidly, and accidents involving containers falling into the sea during voyages are becoming frequent. In harsh environments, container ships are prone to instabilities such as parametric roll, leading to the collapse of large areas of containers. Despite numerous stability optimizations during ship design, stability risks still exist in extreme weather conditions, making it impossible to provide early warning of second-generation stability risks during navigation. Therefore, it is crucial to assess and forecast second-generation intact stability risks in real-world environments by forecasting the marine environment and combining it with the second-generation intact stability standard.

[0003] Based on the practical needs of safe navigation and cargo protection for ships, this paper proposes a second-generation intact stability prediction system for ships operating in real seas. First, the "Second-Generation Intact Stability Standard" is combined with deep learning methods to establish a second-generation stability sensitivity index prediction model. Further model structure optimization and generalization analysis are conducted. Finally, a second-generation intact stability prediction system for ships operating in real seas is developed. This system achieves rapid and accurate regional failure mode sensitivity index prediction, providing conditions for real-time assessment of the second-generation stability criteria for container ships in real seas, thereby ensuring safe navigation of container ships.

[0004] In the current research on ship stability prediction, there has been research on predicting the second-generation stability of ships based on ship parameters. However, the stability of ships sailing in real sea areas changes with changes in the ocean environment and ship navigation parameters. This method ignores the impact of the ocean environment on the second-generation stability assessment of ships, and the accuracy is low. In addition, there are also studies on evaluating the second-generation stability risks of ships in different regions and seasons based on real sea area marine environmental data. Although certain improvements have been made in environmental parameters, this method is only applicable to the ship design stage, and there is still a lack of research on the second-generation stability prediction of ships based on the real sea area marine environment. The present invention focuses on the above problems, establishes a real sea area second-generation stability sensitivity index prediction model, predicts the second-generation stability risk of the ship navigation mission area, and ensures the navigation safety of ships and cargoes.

[0005] Through the above analysis, the problems and defects of the existing technology are as follows:

[0006] Regarding the problem of ship stability prediction, current research on the second-generation stability prediction is mainly carried out from the perspective of ship parameters, ignoring the impact of environmental parameters on stability assessment.

[0007] For the second generation of ship stability research, there are currently methods for improving environmental parameters, but the time scale of environmental parameters is still large, only applicable to ship design stage, and cannot be applied to actual navigation. SUMMARY

[0008] To overcome the problems in the related art, the embodiments of the present application provide a second generation of complete stability prediction system and method for ship real sea navigation, specifically a method for real-time evaluation of ship second generation stability risk through prediction of marine environment data. The purpose of the present application is to establish a real sea area second generation stability sensitivity index prediction model to predict the second generation stability risk of the ship navigation task area and ensure the safety of the ship and cargo navigation. The present application proposes a second generation of complete stability prediction system for ship real sea navigation. The system combines deep learning model with ship second generation stability to build a second generation stability sensitivity index prediction model. The system can realize real-time prediction of ship second generation stability, provide more accurate stability prediction results for route planning, and improve the safety protection capability of the ship in the actual sea area. This technical means is of great significance to ensure the safety of the ship and cargo.

[0009] The technical solution is as follows: a second generation of complete stability prediction method for ship real sea navigation, comprising the following steps:

[0010] S1, acquiring measured marine environment parameters for sea wave prediction and acquiring ship navigation parameters for ship second generation complete stability prediction; the measured marine environment parameters include: wind speed and direction of the sea surface measured by a ship-mounted anemometer, and wave parameters of wave height, wave period and wave direction obtained based on a ship-mounted radar; the ship navigation parameters include: acceleration and angular velocity of ship navigation obtained by an inertial navigation system, and then the heading, speed and position information of ship navigation;

[0011] S2, performing sea wave prediction, and predicting regional significant wave height and average zero-crossing period of sea wave data based on the HP-VIT model;

[0012] S3, using a step-by-step correction algorithm to fuse the collected measured marine environment parameters into the data set of the HP-VIT model for data fusion, inputting the fused data, historical environment data, and combined ship parameter data into a pure stability loss sensitivity index prediction model constructed by using a regional pure stability loss sensitivity index calculation formula to complete pure stability loss sensitivity index prediction; at the same time, inputting the fused data, historical environment data, obtained ship navigation parameters, and combined ship parameter data into a parameter roll sensitivity index prediction model constructed by using a regional parameter roll sensitivity index calculation formula to complete regional parameter roll prediction.

[0013] In step S2, regional significant wave height and average zero-crossing period forecast of ocean wave data is performed based on the HP-VIT model, including:

[0014] By basing the HP-VIT model structure on the VIT model framework, selecting the convolution kernel size and the number of model attention heads, and setting the optimal HP-VIT model structure, the regional significant wave height and average zero-crossing period of ocean wave data are predicted.

[0015] In step S3, the parameter roll sensitivity index prediction model is constructed, including:

[0016] The ship parameter calculation module is used to calculate the average value GM of the ship's main scale and initial stability in waves. m and amplitude GM a The ship parameters are used as input to calculate the roll damping coefficient and The single degree of freedom roll equation is used to calculate and output the different working conditions. value;

[0017] The environmental parameter calculation module uses historical environmental data and measured data from the shipboard measurement system as input, and uses a deep learning-based regional wave forecast model to predict the significant wave height and average zero-crossing period in the future.

[0018] The navigation parameter module uses the speed and heading as inputs, and the ship's navigation parameters are input into the navigation parameter module through the ship's own speedometer and compass measuring instruments;

[0019] Using the sensitivity index prediction module to calculate the ship parameters module The regional wave forecast data of the environmental parameter calculation module, the speed and heading navigation data of the navigation parameter module are used as input; the significant wave height data of the regional wave forecast at each grid position in the navigation sea area within the forecast time is converted into the equivalent regular wave height through the parameter roll equivalent regular wave height calculation formula; then the linear interpolation method is used to obtain the significant wave height data at each grid position in the navigation area under the current ship speed within the forecast time. Results; further determine whether there are risks in various parts of the navigation area, and finally calculate the regional parameter roll sensitivity index of the ship at the current speed; the output of the parameter roll sensitivity index prediction model is the prediction result of the parameter roll sensitivity index at each time and each grid position.

[0020] Furthermore, the parametric roll amplitude is evaluated using the following single-degree-of-freedom roll equation:

[0021]

[0022] Where, I xx +ΔI xxThe sum of the ship's rolling moment of inertia and the additional rolling moment of inertia, in t·m 2 ; is the angular acceleration of the ship's rolling motion, in rad / s 2 ; is the speed of the ship's rolling motion, in rad / s; is the angle of the ship's rolling motion, in rad; δ1 is the damping coefficient of the ship's rolling linear term, δ3 is the damping coefficient of the ship's rolling cubic term; ρ is the density of seawater, in kg / m 3 , take ρ = 1025.0; g is the acceleration due to gravity, unit is m / s 2 , take g = 9.81; The displacement volume for checking loading conditions, in m 3 ; is the righting arm value in the wave at time t, in m; Determine the direction of the restoring moment according to the roll direction to ensure that the moment always points to the horizontal restoration position.

[0023] In the time domain simulation, the fourth-order Runge-Kutta method is used to solve the problem. The initial boundary conditions are

[0024] I xx +ΔI xx The calculation formula is as follows:

[0025]

[0026] Where, T r is the natural period of roll, GM is the initial stability height in still water;

[0027] The calculation formula is as follows:

[0028]

[0029] because yes is an odd function, so The calculation formula is as follows:

[0030]

[0031] GM a =0.5(GM max -GM min )

[0032] ω e =ω-k0V s cos(χ)

[0033]

[0034] k0=2π / λ

[0035] Where, GM max ,GM min are the maximum and minimum values ​​of GM at 10 different crest positions in the wave, ω e ,ω are the ship encounter frequency and wave frequency respectively, k0 is the wave number, V s is the ship speed, χ is the angle between the ship heading and the wave direction, λ is the wavelength, t is the time, a, b, c, d are the GZ random numbers calculated using the least squares method, The fitting coefficients of the changes, is the roll angle.

[0036] Furthermore, the regional parameter roll sensitivity index calculation formula of the ship at the current speed is as follows:

[0037] C2=C S,i ×W i

[0038] Where C2 is the forecast result of the parameter roll sensitivity index at each time and each grid position, W i is a matrix of all 1s, C S,i is the unweighted parametric roll sensitivity index corresponding to the predicted irregular wave sea conditions.

[0039] In step S3, a pure stability loss sensitivity index prediction model is constructed, including:

[0040] The ship parameter calculation module takes the main dimensions of the ship and the righting arm value of the ship in the waves as input, and calculates the heeling arm l PL2 , calculate and output the stability vanishing angle under different working conditions And based on the heel arm l PL2 Heel angle under action

[0041] The environmental parameter calculation module uses historical environmental data as input and uses a deep learning-based regional wave forecast model to predict significant wave heights and average zero-crossing periods in the future.

[0042] The environmental parameter calculation module uses historical wind speed, significant wave height, and average zero-crossing period data as input, and uses a deep learning-based regional wave forecast model to predict significant wave height and average zero-crossing period in the future.

[0043] Using the sensitivity index prediction module to calculate the ship parameter module under different working conditions and The regional wave forecast data of the value and environmental parameter calculation module are used as input. The wave height calculation formula of the equivalent regular wave of pure loss of stability is used to convert the significant wave height data of the regional wave forecast at each grid position in the navigation area within the forecast time into the equivalent regular wave height. Then, the linear interpolation method is used to obtain the wave height of each grid position in the navigation area within the forecast time. and Results; further judge whether there are risks in various parts of the navigation area, the criterion value of pure stability loss C1 i Based on the stability vanishing angle in waves Calculation, criterion value C2 for pure loss of stability i Based on the heel arm l PL2 Heel angle under action Calculation, finally, calculate the regional pure stability loss sensitivity index of the ship. The output of the pure stability loss sensitivity index prediction model is the prediction result of the pure stability loss sensitivity index at each moment and each grid position.

[0044] Furthermore, the criterion value C1 i Based on the stability vanishing angle in the wave The calculation formula is as follows:

[0045]

[0046] Where, is the stability vanishing angle, the minimum value obtained according to the righting arm curve; K PL1 Stability vanishing angle The standard value, K PL1 =30(deg);

[0047] Criterion value C2 i is based on the heeling arm l PL2 Heel angle under action The calculation formula is as follows:

[0048]

[0049] Where,

[0050]

[0051] Where, Fn is the current speed V S The corresponding Froude number is d is the average draft, l PL2 is the heeling arm, H i is the height of the series equivalent regular wave, K PL2 is the heel angle is the standard value of , and λ is the wavelength of the equivalent regular wave.

[0052] Further, the area pure stability loss sensitivity index of the ship is calculated, and the calculation formula is as follows:

[0053] CR1=C1 i * W i

[0054] CR2=C2 i * W i

[0055] In the formula, W i is a full 1 matrix, and the output of the pure stability loss sensitivity index prediction model is the prediction result max (CR1, CR2) of the pure stability loss sensitivity index at each time and each grid position, CR1 is the stability vanishing angle based on the wave The weighted average value is calculated, and CR2 is the heeling angle based on the transverse lever l PL2 under the action of the transverse lever The weighted average value is calculated.

[0056] Another object of the present application is to provide a second-generation complete stability prediction system for real sea navigation of a ship, which implements the second-generation complete stability prediction method for real sea navigation of the ship, and the system comprises:

[0057] A shipborne measurement system provides measured marine environmental parameters and ship navigation parameters for ship dynamic stability calculation; for the measured marine environmental parameters, the shipborne anemograph is used to measure the wind speed and direction on the sea surface to obtain the sea weather conditions, and the wave height, wave period, wave direction and wave parameters are obtained based on the shipborne radar;

[0058] For the ship navigation parameters, the acceleration and angular velocity parameters are obtained through the inertial navigation system, and then the heading, speed and position information are obtained; in this process, the shipborne measurement system collects data in real time to ensure the synchronism and timeliness of the data;

[0059] A sea wave prediction system predicts the regional significant wave height and average zero-crossing period of the sea wave data based on the HP-VIT model;

[0060] The ship second-generation complete stability prediction system fuses the collected measured marine environmental parameters into the data set of the HP-VIT model by using the step-by-step correction algorithm for data fusion, inputs the fused data and historical environmental data, and in combination with the ship parameter data, into a pure stability loss sensitivity index prediction model constructed by using the area pure stability loss sensitivity index calculation formula to complete the pure stability loss sensitivity index prediction; at the same time, the fused data and historical environmental data, the obtained ship navigation parameters, and in combination with the ship parameter data, are input into a parametric roll sensitivity index prediction model constructed by using the area parametric roll sensitivity index calculation formula to complete the area parametric roll prediction.

[0061] Furthermore, the second-generation complete stability prediction system for ships sailing in real sea is carried on a computer-readable storage medium, and the computer-readable storage medium stores a computer program. When the computer program is executed by the processor, it can realize the functions of the second-generation complete stability prediction system for ships sailing in real sea.

[0062] In combination with all the above technical solutions, the beneficial effects of the present invention are as follows:

[0063] First, the present invention establishes a failure mode sensitivity index prediction model to enable rapid prediction of ship second-generation stability in real-sea environments. By improving the calculation method of the "Second-Generation Intact Stability Criterion" based on deep learning methods, a prediction model for parametric roll and pure loss of stability sensitivity index was established. This model was initially validated using data from an accident ship, enabling high-precision prediction of ship second-generation stability in real-sea environments.

[0064] Second, the present invention combines the "Second Generation Intact Stability Standard" with deep learning methods, conducts model structure optimization analysis and model generalization analysis, and establishes a high-precision prediction model for the regional failure mode sensitivity index. Further, taking a typical container accident ship as the research target, the stability criterion assessment results of the accident ship are calculated and the accuracy of the sensitivity index prediction results is verified. Compared with the traditional evaluation method, the second-generation stability prediction model proposed in the present invention solves the problem that the traditional IMO method is prone to underestimate the actual risk, and can realize the rapid prediction of the second-generation stability of ships in real sea environments. The research results can provide auxiliary decision-making information for the navigation safety of container ships, route planning corrections, etc., which is of great significance to ensuring the safety of ships and cargoes.

[0065] Third, the present invention significantly improves the accuracy and timeliness of ship navigation assessments, enhances ship hazard warning and emergency response capabilities, reduces economic and social losses, and promotes the development of intelligent shipping. By combining stability assessment standards and deep learning methods to construct a high-precision prediction system for regional failure mode sensitivity index, it is possible to achieve real-time assessment and prediction of the second-generation stability of ships in actual sea areas, significantly shorten the forecast time, improve the accuracy of assessment and forecasting, and provide real-time and reliable forecast information for government decision-making departments, emergency management agencies and marine-related industries. At the same time, the system is lightweight and easy to deploy, and is suitable for government departments, scientific research institutions and corporate users of different sizes. The system can be expanded and applied to the global navigation safety assurance and route planning correction service market, and has broad market promotion potential; it can also cooperate with ship insurance institutions to provide disaster probability analysis and risk assessment data, optimize insurance product pricing, and create data value-added benefits.

[0066] Fourthly, the application is directed to the problem that the traditional second-generation stability evaluation ignores the difference of the real sea area marine environment and cannot accurately evaluate the navigation risk. The application combines the Second Generation Complete Stability Standard with the deep learning method and proposes an innovative solution to realize high-precision prediction of the regional failure mode sensitivity index, breaks through the limitation of marine environment data precision on the accuracy of stability evaluation, well solves the problem that the traditional IMO method easily underestimates the actual risk, makes up for the deficiency of the standard which is only used for stability evaluation in the dynamic design stage of the ship, innovatively applies the Second Generation Complete Stability Standard to the navigation stage of the ship, provides auxiliary decision-making information for the navigation of the container ship, and has great significance for the safety of the ship and the cargo.

[0067] Fifthly, the application is directed to the problem that the traditional second-generation stability evaluation scheme relies on long-term environmental statistical data and cannot reflect the regional difference of the real sea area, resulting in deviation of risk evaluation. The application is based on data-driven real sea area environment modeling and uses the deep learning method to replace physical simplification, so as to get rid of the dependence of the traditional stability evaluation method on long-term environmental statistical data. In addition, the traditional standard only provides static threshold judgment and lacks real-time evaluation of dynamic failure modes. The application proposes a HP-VIT model based on deep learning, which greatly shortens the prediction time of the model while improving the calculation precision and has strong generalization ability. The wave data predicted by the model is input into the improved stability evaluation module in real time to realize real-time high-precision prediction of the failure mode sensitivity index of the ship during navigation. This technical breakthrough is an effective improvement and innovation of the current traditional second-generation stability static evaluation scheme.

[0068] Sixthly, the application applies the Second Generation Complete Stability Standard to the navigation stage of the ship to realize real-time evaluation and prediction of the stability of the ship. In the prior art, the second-generation stability standard is only used in the design stage of the ship and the accuracy is limited by the environmental data. The use of long-term environmental statistical data easily underestimates the actual risk. The application combines the standard with the deep learning method and uses the environmental data predicted by the deep learning model as the input of the stability calculation, so that more accurate and higher resolution marine environment data can be obtained, and the occurrence position and time of the risk can be more accurately identified. The application well solves the problem that the existing method easily underestimates the actual risk and can be used in the navigation stage of the ship to realize rapid prediction of the second-generation stability of the ship under the real sea environment. BRIEF DESCRIPTION OF DRAWINGS

[0069] The accompanying drawings, which are incorporated into and form a part of the specification, illustrate embodiments consistent with the present disclosure and, together with the description, serve to explain the principles of the disclosure;

[0070] Figure 1 is a second-generation complete stability prediction method flowchart for ship real sea navigation provided by the embodiment of the application;

[0071] Figure 2 1 is a schematic diagram of a parameter roll sensitivity index prediction model provided by an embodiment of the present invention;

[0072] Figure 3 Schematic diagram of a pure stability loss sensitivity index prediction model provided by an embodiment of the present invention;

[0073] Figure 4 This is a structural diagram of the HP-VIT model provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0074] To make the above-mentioned objects, features, and advantages of the present invention more readily apparent, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings. The following description sets forth numerous specific details to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art may make similar modifications without departing from the scope of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0075] To address the problem that traditional second-generation stability assessments ignore the variability of real-world marine environments and fail to accurately assess navigation risks, the present invention innovates by combining the Second-Generation Intact Stability Standard with deep learning methods to propose a deep-learning-based HP-VIT model. This model enables intelligent forecasting of regional significant wave heights and average zero-crossing periods, significantly reducing prediction time while improving computational accuracy. Furthermore, by combining and improving the regional wave forecast model with traditional second-generation stability assessment methods, a deep-learning-based failure mode sensitivity index prediction model is developed, paving the way for real-time second-generation stability assessments in real-world environments.

[0076] Example 1. The second-generation complete stability prediction method for ships sailing in real seas provided by the embodiment of the present invention includes: ship-borne measurement, which provides real-time environmental measurement data and ship navigation parameters for subsequent stability prediction calculations; wave forecast, which provides future marine environmental information of the target sea area for ship stability prediction; the second-generation complete stability prediction of ships, which provides future stability risk information of ships in the target sea area by combining deep learning methods with second-generation stability assessment methods, thereby realizing second-generation stability prediction of ships in real sea areas.

[0077] like Figure 1 As shown, specifically including:

[0078] S1, obtaining measured ocean environmental parameters for wave forecasting and obtaining ship navigation parameters for second-generation intact stability prediction of ships; the measured ocean environmental parameters include: wind speed and direction on the ocean surface measured by a ship-borne anemometer, and wave parameters such as wave height, wave period, and wave direction obtained based on a ship-borne radar; the ship navigation parameters include: acceleration and angular velocity of the ship obtained by an inertial navigation system, and further obtained ship heading, speed, and position information;

[0079] S2, conduct ocean wave forecast, and forecast the regional significant wave height and average zero-crossing period of ocean wave data based on the HP-VIT model;

[0080] S3, using the stepwise correction algorithm to integrate the collected measured ocean environmental parameters into the data set of the HP-VIT model for data fusion, the fused data and historical environmental data, as well as the ship parameter data, are input into the regional pure stability loss sensitivity index calculation formula to construct the pure stability loss sensitivity index prediction model (such as Figure 3 At the same time, the fused data, historical environmental data, obtained ship navigation parameters, and combined with ship parameter data are input into the parameter roll sensitivity index prediction model constructed using the regional parameter roll sensitivity index calculation formula (as shown in Figure 2 As shown), regional parameter roll forecast is completed.

[0081] Exemplarily, step S1 provides measured ocean environment parameters for the wave forecast in step S2, and provides ship navigation parameters for the second-generation intact stability forecast of the ship in step S3.

[0082] Exemplarily, in step S2, wave forecasting is implemented using a deep learning model. Based on the measured ocean environmental parameters obtained in step S1, wave data is forecasted using the HP-VIT model. This is based on the HP-VIT (Vision Transformer) model, which integrates an attention mechanism and convolution operations. The HP-VIT model structure is optimized and analyzed, and parameters such as the convolution kernel size and the number of attention heads are optimized. Further, the HP-VIT model generalization analysis is performed to test the HP-VIT model's forecasting capabilities in different sea areas. Finally, the accuracy of the HP-VIT model is verified using the root mean square error (RMSE), mean absolute percentage error (MAPE), and mean absolute error (MAE) evaluation indicators to complete the regional significant wave height and average zero-crossing period forecast for the wave data.

[0083] Exemplarily, the second-generation complete stability prediction system for real sea navigation of a ship provided by the embodiment of the present application comprises:

[0084] The shipborne measurement system provides measured sea environment parameters and ship navigation parameters for ship dynamic stability calculation. For the measured sea environment parameters, the shipborne anemograph is used to measure the wind speed and direction on the sea surface to obtain the sea weather conditions, and the shipborne radar is used to obtain wave parameters such as wave height, wave period and wave direction.

[0085] For the ship navigation parameters, the inertial navigation system is used to obtain parameters such as acceleration and angular velocity, and then to obtain information such as heading, speed and position. In this process, the shipborne measurement system collects data in real time to ensure the synchronicity and timeliness of the data.

[0086] The sea wave prediction system performs high-precision prediction of sea wave data based on the HP-VIT model, performs structural optimization analysis of the HP-VIT model, and performs parameter optimization on the convolution kernel size and the number of attention heads. Further, the HP-VIT model is analyzed for generalization to test the prediction ability of the HP-VIT model in different sea areas. Finally, the precision of the HP-VIT model is verified by using evaluation indexes such as Root of Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and Mean Absolute Error (MAE), and the regional significant wave height and average zero-crossing period prediction of sea wave data is completed.

[0087] The ship second-generation complete stability prediction system uses the step-by-step correction algorithm to fuse the collected measured sea environment parameters into the data set of the HP-VIT model for data fusion, improves the input data precision of the HP-VIT model, and provides environmental parameters for the second-generation stability prediction of the ship. The fused data and historical environmental data, combined with ship parameter data, are jointly input into a pure stability loss sensitivity index prediction model (such as Figure 3 ) constructed by using the regional pure stability loss sensitivity index calculation formula to complete the pure stability loss sensitivity index prediction; at the same time, the fused data and historical environmental data, obtained ship navigation parameters, combined with ship parameter data, are jointly input into a parameter roll sensitivity index prediction model (such as Figure 2 ) constructed by using the regional parameter roll sensitivity index calculation formula to complete the regional parameter roll prediction. This realizes rapid and accurate regional failure mode sensitivity index prediction, provides conditions for real-time evaluation of the second-generation stability criterion of a container ship in a real sea area, and ensures the navigation safety of the container ship.

[0088] For example, in step S1, the shipborne measurement system obtains information such as wind speed and direction, wave height, wave period, wave direction, ship heading, speed, and position on the ocean surface, and then uses the measured information as input to the ship navigation parameter module and the environmental parameter calculation module (e.g. Figure 2 、 Figure 3 ).

[0089] For example, in step S2, the ocean wave forecasting system proposes a deep learning-based HP-VIT model, which is an improvement based on the VIT model framework and can realize intelligent forecasting of regional significant wave height and average zero-crossing period. The HP-VIT model calculates the forecast results of regional significant wave height and average zero-crossing period, and then uses the output of the HP-VIT model as the input of the environmental parameter calculation module of the sensitivity index forecasting model (such as Figure 2 、 Figure 3 ).

[0090] like Figure 4 As shown in Figure 1, the HP-VIT model is based on the VIT model framework and integrates the attention mechanism and convolution operations for data encoding and decoding.

[0091] First, input data, including wind speed, significant wave height, and average zero-crossing period, is preprocessed. The model's convolutional layers extract features from image patches, effectively capturing local information within the image through sliding operations on the convolution kernel. Subsequently, positional encoding is used to incorporate sequential information about elements in the data, enhancing the model's understanding of time series. Because the model's input ocean environment data is complex, the model may learn noisy or unimportant features. Therefore, regularization techniques are employed to prevent overfitting.

[0092] Next, the regularized data is fed into the model encoding module. These features are further processed and encoded for model analysis. To further stabilize the training process, normalization techniques are applied to the encoding block to standardize the layer inputs, accelerating learning and improving model stability. The multi-head attention mechanism, a key component of the encoding block, enhances the model's ability to capture complex dependencies by allowing it to process information in multiple subspaces in parallel. Following the multi-head attention process, regularization and normalization steps are applied again to maintain network training stability and help the model learn more robust feature representations.

[0093] The normalized data is then fed into a multilayer perceptron (MLP) block, which consists of multiple fully connected layers. This structure uses nonlinear activation functions to introduce nonlinear capabilities to the network, thereby capturing complex relationships in the input data. To improve the model's generalization and prevent overfitting, regularization techniques are applied after the MLP. Subsequently, multiple linear layers further extract and transform the data, while the reapplied regularization step ensures the model's stability during training.

[0094] Finally, the processed data is reshaped into the format required for predicting significant wave height and average zero-crossing period, and further transposed convolution is performed to adjust the size of the feature map and output it.

[0095] Exemplarily, in step S2, the HP-VIT model structure optimization analysis refers to analyzing the convolution kernel size and the number of model attention heads in the model to optimize the HP-VIT model structure. By setting different convolution kernel sizes and the number of model attention heads, the impact of the convolution kernel size and the number of attention heads on the model prediction accuracy is analyzed, and the optimal convolution kernel size and the number of model attention heads for the HP-VIT model are further selected to set the optimal HP-VIT model structure.

[0096] Exemplarily, in step S2, the HP-VIT model generalization analysis refers to analyzing the sea area adaptability of the model, verifying the robustness and reliability of the model in different environments, and verifying the sea area generalization of the system.

[0097] Exemplarily, in step S3, the data fusion method adopts a step-by-step correction algorithm. This algorithm integrates the measured ocean environmental parameters collected by the shipborne measurement system into the data set of the HP-VIT model, thereby improving the accuracy of the environmental parameter forecast by improving the input data accuracy of the HP-VIT model, thereby improving the accuracy of the environmental input of the sensitivity index forecast model and the accuracy of the stability risk forecast. The algorithm gradually improves the accuracy of the integrated data by gradually correcting each data source. Assume that there are N data sources to be integrated, namely x1, x2...x N , the result of integration is y. The weight of each data source is ω i , all weights are initially 1.

[0098] Exemplarily, the steps of the step-by-step correction algorithm are as follows:

[0099] Data preprocessing: Perform preprocessing operations such as cleaning, denoising, and normalization on the data of each data source to ensure data consistency and comparability.

[0100] Initial fusion: Perform preliminary fusion of data from multiple data sources to obtain an initial data set.

[0101] Calculate weights: For each target point, calculate the weights of all data points in each data source. The weights are calculated using the Cressman method, and the formula is as follows:

[0102]

[0103] Where, ω i is the weight of the i-th data point, r0 is the parameter that controls the fusion radius, d i is the distance between the i-th data point and the target point.

[0104] Calculate the weighted average: For each target point, calculate the weighted average based on the weight of each data source. The formula is as follows:

[0105]

[0106] In the formula, y is the target point value after fusion, y i is the value of the i-th data point, ω i is the weight of the i-th data point, and n is the number of data sources.

[0107] Model evaluation: Evaluate and test the trained model to determine the accuracy and precision of the model.

[0108] Correction and update: Based on the evaluation results of the model, the dataset is corrected and updated to improve the accuracy and precision of the model.

[0109] Repeated training and updating: Repeat the steps of model training, evaluation, and correction and updating until the model converges and obtains satisfactory results.

[0110] For example, the parameter roll sensitivity index prediction model in step S3 is innovatively proposed in the present invention and constructed as follows: the parameter roll sensitivity index prediction model based on deep learning is divided into four modules: ship parameter calculation module, environmental parameter calculation module, navigation parameter module, and sensitivity index prediction module. The specific calculation method and process of each module are shown in the attached Figure 2 .

[0111] The ship parameter calculation module is used to calculate the average value GM of the ship's main scale and initial stability in waves. m and amplitude GM a The ship parameters such as the roll damping coefficient and the The single degree of freedom roll equation is used to calculate and output the different working conditions. The parametric roll amplitude should be evaluated using the following single-degree-of-freedom roll equation:

[0112]

[0113] Where, Ixx +ΔI xx The sum of the ship's rolling moment of inertia and the additional rolling moment of inertia, in t·m 2 ; is the angular acceleration of the ship's rolling motion, in rad / s 2 ; is the speed of the ship's rolling motion, in rad / s; is the angle of the ship's rolling motion, in rad; δ1 is the damping coefficient of the ship's rolling linear term, δ3 is the damping coefficient of the ship's rolling cubic term; ρ is the density of seawater, in kg / m 3 , take ρ = 1025.0; g is the acceleration due to gravity, unit is m / s 2 , take g = 9.81; The displacement volume for checking loading conditions, in m 3 ; is the value of the righting arm in the wave at time t, in m; Determine the direction of the restoring moment according to the roll direction to ensure that the moment always points to the horizontal restoration position.

[0114] In the time domain simulation, the fourth-order Runge-Kutta method is used to solve the problem. The initial boundary conditions are

[0115] I xx +ΔI xx The calculation formula is as follows:

[0116]

[0117] Where, T r is the natural period of roll, GM is the initial stability height in still water;

[0118] The calculation formula is as follows:

[0119]

[0120] because yes is an odd function, so The calculation formula is as follows:

[0121]

[0122] GM a =0.5(GM max -GM max )

[0123] ω e =ω-k0V s cos(χ)

[0124]

[0125] k0=2π / λ

[0126] Where, GM max ,GM min are the maximum and minimum values ​​of GM at 10 different crest positions in the wave, ω e ,ω are the ship encounter frequency and wave frequency respectively, k0 is the wave number, V s is the ship speed, χ is the angle between the ship heading and the wave direction, λ is the wavelength, t is the time, a, b, c, d are the GZ random numbers calculated using the least squares method, The fitting coefficients of the changes, is the roll angle.

[0127] The environmental parameter calculation module uses historical environmental data and measured data from the shipborne measurement system as input, and uses a regional wave forecasting model based on deep learning to predict the significant wave height and average zero-crossing period at future times.

[0128] The data processing flow here is: wind speed (u 10 ,v 10 ), significant wave height (Hs) and average zero-crossing period (T 02 ) The historical environmental data and the measured data from the shipborne measurement system are gradually corrected to obtain the input data for the regional wave forecast model (i.e., the HP-VIT model). This data is processed by the HP-VIT model to realize the forecast of significant wave heights and average zero-crossing periods at future moments (i.e., the historical environmental data are gradually corrected using the measured data, and the corrected data are then substituted into the HP-VIT model for environmental forecasting).

[0129] Therefore, the original input of the environmental parameter calculation module is the historical environmental data and the measured data of the shipborne measurement system, which are unprocessed data.

[0130] The environmental parameter calculation module takes historical environmental data and measured data from the shipborne measurement system as input, uses the measured data to gradually revise the historical data to obtain more accurate regional environmental data, and substitutes this data into the regional wave forecast model - the HP-VIT model. After the data preprocessing module, encoding module, multi-layer perceptron block and decoding module, the forecast of significant wave height and average zero-crossing period at future moments is realized.

[0131] The navigation parameter module takes speed and heading as input, and inputs the ship's navigation parameters into the navigation parameter module through the ship's own measuring instruments such as speedometer and compass.

[0132] The sensitivity index prediction module is based on the ship parameter calculation module The inputs are the regional wave forecast data of the value, environmental parameter calculation module, and navigation data such as speed and heading of the navigation parameter module. First, the significant wave height data of the regional wave forecast at each grid position in the navigation area within the forecast time is converted into the equivalent regular wave height through the parameter roll equivalent regular wave height calculation formula. Then, the linear interpolation method is used to obtain the significant wave height data at each grid position in the navigation area within the forecast time at the current ship speed. Further judge whether there are risks in the navigation area, the maximum roll amplitude of the parameter roll If it is greater than or equal to 25°, then take C s,i =1, otherwise take C s,i = 0. Finally, the present invention innovatively proposes to calculate the regional parameter roll sensitivity index of the ship at the current speed, and the calculation formula is as follows:

[0133] C2=C s,i ×W i

[0134] Where C2 is the forecast result of the parameter roll sensitivity index at each time and each grid position, W i is a matrix of all 1s, C S,i is the unweighted parametric roll sensitivity index corresponding to the predicted irregular wave sea conditions.

[0135] The final output of the parameter roll sensitivity index prediction model is the prediction result C2 of the parameter roll sensitivity index at each time and each grid position.

[0136] Taking a sea area with a wave forecast accuracy of 0.25°×0.25°, a grid number of 10×10, a forecast period of 6 hours, and a time resolution of 1 hour as an example, the calculated equivalent regular wave height is a matrix of 10×10×6, and the maximum roll amplitude of the parameter roll is C s,i The results of C2 and C3 are both 10 × 10 × 6 matrices. This means that at each moment, each grid position has a predicted result for the parametric roll sensitivity index. Traditional assessment methods only provide a single fixed assessment result for each ship, ignoring the effects of time and space. Compared to traditional assessment methods, this model significantly improves the accuracy of parametric roll risk assessment.

[0137] For example, the pure stability loss sensitivity index prediction model in step S3 is innovatively proposed in the present invention and constructed as follows: the pure stability loss sensitivity index prediction model based on deep learning is divided into three modules: ship parameter calculation module, environmental parameter calculation module, and sensitivity index prediction module. The specific calculation method and process of each module are shown in the attached Figure 3 .

[0138] The ship parameter calculation module takes the main dimensions of the ship and the righting arm value of the ship in the waves as input, and calculates the heeling arm l PL2 , calculate and output the stability vanishing angle under different working conditions And based on the heel arm l PL2 Heel angle under action

[0139] Criterion value C1 i Based on the stability vanishing angle in the wave The calculation formula is as follows:

[0140]

[0141] Where, is the stability vanishing angle, the minimum value obtained according to the righting arm curve; K PL1 Stability vanishing angle The standard value, K PL1 =30(deg);

[0142] Criterion value C2 i is based on the heeling arm l PL2 Heel angle under action The calculation formula is as follows:

[0143]

[0144] Where,

[0145]

[0146] Where Fn is the current speed V S The corresponding Froude number is d is the average draft, l PL2 is the heeling arm, H i is the height of the series equivalent regular wave, K PL2 is the heel angle is the standard value of , and λ is the wavelength of the equivalent regular wave.

[0147] The environmental parameter calculation module uses historical environmental data as input and uses a deep learning-based regional wave forecast model to predict significant wave heights and average zero-crossing periods in the future. The regional wave forecast model is the HP-VIT model mentioned above. This model predicts future wave data in the region based on the region's historical and measured environmental data. The forecast results serve as environmental input for stability assessment.

[0148] The sensitivity index prediction module is based on the different working conditions of the ship parameter calculation module. and The regional wave forecast data of the calculation module of the value and environmental parameters are used as input. First, the wave height calculation formula of the equivalent regular wave of pure loss of stability is used to convert the significant wave height data of the regional wave forecast at each grid position in the navigation area within the forecast time into the equivalent regular wave height. Then, the linear interpolation method is used to obtain the wave height of each grid position in the navigation area within the forecast time. and Further judge whether there are risks in various parts of the navigation area, the criterion value of pure stability loss C1 i Based on the stability vanishing angle in the wave Calculated, If it is less than 30°, take C1 i =1, otherwise take C1 i = 0. Criterion value of pure loss of stability C2 i is based on the heeling arm l PL2 Heel angle under action Calculated, for container ships, If it is greater than 25°, then take C2 i =1, the meaning of this parameter has been described in the previous line, which is the criterion value for loss of initial stability; in other cases, C2 i =0. Finally, the present invention innovatively proposes to calculate the regional pure stability loss sensitivity index of the ship, and the calculation formula is as follows:

[0149] CR1=C1 i ×W i

[0150] CR2=C2 i ×W i

[0151] Where W i The output of the pure loss of stability sensitivity index prediction model is the prediction result of the pure loss of stability sensitivity index at each moment and each grid position max(CR1,CR2), where CR1 is the stability loss angle based on the wave. The weighted criterion value calculated is CR2, which is based on the heeling lever l PL2 Heel angle under action Calculated weighted criterion value.

[0152] The final output of the pure loss of stability sensitivity index prediction model is the prediction result max(CR1,CR2) of the pure loss of stability sensitivity index at each time and each grid position.

[0153] Taking a sea area with a wave forecast accuracy of 0.25°×0.25°, a grid number of 10×10, a forecast period of 6 hours, and a time resolution of 1 hour as an example, the calculated equivalent regular wave height is a matrix of 10×10×6, and the pure stability loss criterion value C1 i、C2 i The results for CR1, CR2, and max(CR1, CR2) are all 10 × 10 × 6 matrices. This means that at every moment, each grid location has a prediction result for the pure loss of stability sensitivity index. Traditional assessment methods only provide a single fixed assessment result for each ship, without considering the effects of time and space. Compared to traditional assessment methods, this model significantly improves the accuracy of pure loss of stability risk assessment.

[0154] It can be seen from the above embodiments that the present invention adopts two existing technologies (wave prediction technology based on deep learning and second-generation intact stability criterion assessment technology), integrates the two technological innovations, improves the calculation formulas and methods therein, and proposes a new second-generation stability prediction model, which breaks through and solves the second-generation stability prediction problem.

[0155] There are five second-generation stability failure modes, including parametric roll, pure loss of stability, overacceleration, wave riding / rolling, and dead ship. The present invention constructs a sensitivity index prediction model for the two failure modes of parametric roll and pure loss of stability, but the prediction method proposed in the present invention is also applicable to the other three modes.

[0156] This invention improves upon the traditional second-generation stability assessment method. The environmental parameters of the traditional IMO assessment method use North Atlantic environmental data, which is single and based on long-term statistical analysis, and lacks real-time performance. By combining the "Second Generation Intact Stability Standard" with deep learning methods, this invention establishes a high-precision prediction model for the regional failure mode sensitivity index with higher spatiotemporal resolution and improved accuracy, enabling more accurate identification of the location and timing of risk occurrence.

[0157] The present invention provides a method for inputting the results of measured data collection and regional wave forecast into a sensitivity index forecast model.

[0158] Existing technologies cannot achieve regional stability risk forecasts, and even single-point forecasts are only very short-term or short-term forecasts. However, the present invention can achieve regional, 24-hour stability risk forecasts.

[0159] Experiment 1: A container anomaly of a certain Dyros ship occurred in the NW3 sea area of ​​a certain ocean. However, in the annual and four-season assessments of the Dyros ship in the NW3 sea area, there was no risk of parametric roll or pure loss of stability. Based on the failure mode sensitivity index prediction model based on deep learning constructed by the present invention, the parametric roll and pure loss of stability sensitivity index of the sea area and the Dyros ship during the abnormal period are predicted. The parameter roll risk that the sea area is prone to can be obtained through the prediction result of the parameter roll sensitivity index of the Dyros ship during the abnormal period. At the same time, the pure loss of stability sensitivity index prediction result of the Dyros ship at the abnormal moment is 0, and there is no pure loss of stability risk. It can be seen from the results that there is indeed a parameter roll risk in the NW3 sea area during the abnormal period. This conclusion is consistent with the actual situation, indicating that the model proposed by the present invention can more accurately predict the parameter roll risk.

[0160] Experiment 2: An abnormality occurred in the container of a certain Essen ship in the NE28 sea area of ​​a certain ocean. However, in the annual and four-season assessments of the Essen ship in the NE28 sea area, there was no risk of parametric roll or pure loss of stability. Based on the failure mode sensitivity index prediction model based on deep learning constructed by the present invention, the parametric roll and pure loss of stability sensitivity index of the sea area and the Essen ship during the abnormal period are predicted. Through the prediction results of the parameter roll sensitivity index of the Essen ship during the abnormal period, it can be known that a certain sea area is prone to the risk of parameter roll. At the same time, the prediction result of the pure loss of stability sensitivity index of the Essen ship at the abnormal moment is 0, and there is no pure loss of stability risk. It can be seen from the results that there is indeed a risk of parameter roll in the NE28 sea area during the abnormal period, and this conclusion is consistent with the actual situation. The calculation results of the present invention are consistent with the conclusions of the report, indicating that the model proposed by the present invention can more accurately predict the risk of parameter roll.

[0161] The above description is only a preferred specific implementation method of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions and improvements made by any technician familiar with this technical field within the technical scope disclosed by the present invention and within the spirit and principles of the present invention should be covered by the scope of protection of the present invention.

Claims

1. A second-generation intact stability prediction method for ships sailing in real sea conditions, characterized by: The method comprises the following steps: S1, obtaining measured ocean environmental parameters for wave forecasting and obtaining ship navigation parameters for second-generation intact stability prediction of ships; the measured ocean environmental parameters include: wind speed and direction on the ocean surface measured by a ship-borne anemometer, and wave parameters such as wave height, wave period, and wave direction obtained based on a ship-borne radar; the ship navigation parameters include: acceleration and angular velocity of the ship obtained by an inertial navigation system, and further obtained ship heading, speed, and position information; S2, conduct ocean wave forecast, and forecast the regional significant wave height and average zero-crossing period of ocean wave data based on the HP-VIT model; S3, using the step-by-step correction algorithm to fuse the collected measured ocean environmental parameters into the data set of the HP-VIT model for data fusion, the fused data are input together with the historical environmental data, and combined with the ship parameter data, into the pure stability loss sensitivity index prediction model constructed using the regional pure stability loss sensitivity index calculation formula to complete the pure stability loss sensitivity index prediction; at the same time, the fused data are input together with the historical environmental data, the obtained ship navigation parameters, and combined with the ship parameter data into the parameter roll sensitivity index prediction model constructed using the regional parameter roll sensitivity index calculation formula to complete the regional parameter roll prediction.

2. The second-generation intact stability prediction method for ships sailing in real sea conditions according to claim 1 is characterized in that: In step S2, regional significant wave height and average zero-crossing period forecast of ocean wave data is performed based on the HP-VIT model, including: By basing the HP-VIT model structure on the VIT model framework, selecting the convolution kernel size and the number of model attention heads, and setting the optimal HP-VIT model structure, the regional significant wave height and average zero-crossing period of ocean wave data are predicted.

3. The second-generation intact stability prediction method for ships sailing in real sea conditions according to claim 1 is characterized in that: In step S3, the parameter roll sensitivity index prediction model is constructed, including: The ship parameter calculation module is used to calculate the average value GM of the ship's main scale and initial stability in waves. m and amplitude GM a The ship parameters are used as input to calculate the roll damping coefficient and The single degree of freedom roll equation is used to calculate and output the different working conditions. value; The environmental parameter calculation module uses historical environmental data and measured data from the shipboard measurement system as input, and uses a deep learning-based regional wave forecast model to predict the significant wave height and average zero-crossing period in the future. The navigation parameter module uses the speed and heading as inputs, and the ship's navigation parameters are input into the navigation parameter module through the ship's own speedometer and compass measuring instruments; Using the sensitivity index prediction module to calculate the ship parameters module The regional wave forecast data of the environmental parameter calculation module, the speed and heading navigation data of the navigation parameter module are used as input; the significant wave height data of the regional wave forecast at each grid position in the navigation sea area within the forecast time is converted into the equivalent regular wave height through the parameter roll equivalent regular wave height calculation formula; then the linear interpolation method is used to obtain the significant wave height data at each grid position in the navigation area under the current ship speed within the forecast time. Results; further determine whether there are risks in various parts of the navigation area, and finally calculate the regional parameter roll sensitivity index of the ship at the current speed; the output of the parameter roll sensitivity index prediction model is the prediction result of the parameter roll sensitivity index at each time and each grid position.

4. The second-generation intact stability prediction method for ships sailing in real sea conditions according to claim 3 is characterized in that: The parametric roll amplitude is evaluated using the following single-degree-of-freedom roll equation: Where, I xx +ΔI xx The sum of the ship's rolling moment of inertia and the additional rolling moment of inertia, in t·m 2 ; is the angular acceleration of the ship's rolling motion, in rad / s 2 ; is the speed of the ship's rolling motion, in rad / s; is the angle of the ship's rolling motion, in rad; δ1 is the damping coefficient of the ship's rolling linear term, δ3 is the damping coefficient of the ship's rolling cubic term; ρ is the density of seawater, in kg / m 3 , take ρ = 1025.0; g is the acceleration due to gravity, unit is m / s 2 , take g = 9.81; The displacement volume for checking loading conditions, in m 3 ; is the righting arm value in the wave at time t, in m; The direction of the restoring torque is determined according to the roll direction to ensure that the torque always points to the horizontal recovery position; in the time domain simulation, the fourth-order Runge-Kutta method is used to solve the initial boundary conditions. I xx +ΔI xx The calculation formula is as follows: Where, T r is the natural period of roll, GM is the initial stability height in still water; The calculation formula is as follows: because yes is an odd function, so The calculation formula is as follows: GM a =0.5(GM max -GM min ) oh e =ω-k0V s cos(x) k0=2π / λ Where, GM max ,GM min are the maximum and minimum values ​​of GM at 10 different crest positions in the wave, ω e ,ω are the ship encounter frequency and wave frequency respectively, k0 is the wave number, V s is the ship speed, χ is the angle between the ship heading and the wave direction, λ is the wavelength, t is the time, a, b, c, d are the GZ random numbers calculated using the least squares method, The fitting coefficients of the changes, is the roll angle.

5. The second-generation intact stability prediction method for ships sailing in real sea conditions according to claim 3 is characterized in that: The calculation formula of the regional parameter roll sensitivity index of the ship at the current speed is as follows: C2=C S,i ×W i Where C2 is the forecast result of the parameter roll sensitivity index at each time and each grid position, W i is a matrix of all 1s, C S,i is the unweighted parametric roll sensitivity index corresponding to the predicted irregular wave sea conditions.

6. The second-generation intact stability prediction method for ships sailing in real sea conditions according to claim 1 is characterized in that: In step S3, a pure stability loss sensitivity index prediction model is constructed, including: The ship parameter calculation module takes the main dimensions of the ship and the righting arm value of the ship in the waves as input, and calculates the heeling arm l PL2 , calculate and output the stability vanishing angle under different working conditions And based on the heel arm l PL2 Heel angle under action The environmental parameter calculation module uses historical environmental data as input and uses a deep learning-based regional wave forecast model to predict significant wave heights and average zero-crossing periods in the future. The environmental parameter calculation module uses historical wind speed, significant wave height, and average zero-crossing period data as input, and uses a deep learning-based regional wave forecast model to predict significant wave height and average zero-crossing period in the future. Using the sensitivity index prediction module to calculate the ship parameter module under different working conditions and The regional wave forecast data of the value and environmental parameter calculation module are used as input. The wave height calculation formula of the equivalent regular wave of pure loss of stability is used to convert the significant wave height data of the regional wave forecast at each grid position in the navigation area within the forecast time into the equivalent regular wave height. Then, the linear interpolation method is used to obtain the wave height of each grid position in the navigation area within the forecast time. and Results; further judge whether there are risks in various parts of the navigation area, the criterion value of pure stability loss C1 i Based on the stability vanishing angle in waves Calculation, criterion value C2 for pure loss of stability i Based on the heel arm l PL2 Heel angle under action Calculation, finally, calculate the regional pure stability loss sensitivity index of the ship. The output of the pure stability loss sensitivity index prediction model is the prediction result of the pure stability loss sensitivity index at each moment and each grid position.

7. The second-generation intact stability prediction method for ships sailing in real sea conditions according to claim 6 is characterized in that: Criterion value C1 i Based on the stability vanishing angle in the wave The calculation formula is as follows: Where, is the stability vanishing angle, the minimum value obtained according to the righting arm curve; K PL1 Stability vanishing angle The standard value, K PL1 =30(deg); Criterion value C2 i is based on the heeling arm l PL2 Heel angle under action The calculation formula is as follows: Where, Where Fn is the current speed V S The corresponding Froude number is d is the average draft, l PL2 is the heeling arm, H i is the height of the series equivalent regular wave, K PL2 is the heel angle is the standard value of , and λ is the wavelength of the equivalent regular wave.

8. The second-generation intact stability prediction method for ships sailing in real sea conditions according to claim 6 is characterized in that: Calculate the regional pure stability loss sensitivity index of the ship, the calculation formula is as follows: CR1=C1 i ×W i CR2=C2 i ×W i Where W i The output of the pure loss of stability sensitivity index prediction model is the prediction result of the pure loss of stability sensitivity index at each moment and each grid position max(CR1,CR2), where CR1 is the stability loss angle based on the wave. The weighted criterion value calculated is CR2, which is based on the heeling lever l PL2 Heel angle under action Calculated weighted criterion value.

9. A second-generation intact stability prediction system for ships sailing in real sea conditions, characterized by: The system implements the second-generation intact stability prediction method for ships sailing in real seas as described in any one of claims 1 to 8, and the system includes: The shipborne measurement system provides measured ocean environmental parameters and navigation parameters for ship dynamic stability calculations. Based on the measured ocean environmental parameters, the shipborne anemometer measures the wind speed and direction on the ocean surface to obtain marine meteorological conditions, and the shipborne radar is used to obtain wave parameters such as wave height, wave period, and wave direction. For ship navigation parameters, the inertial navigation system obtains acceleration and angular velocity parameters, and then obtains heading, speed, and position information. During this process, the shipboard measurement system collects data in real time to ensure data synchronization and timeliness. The ocean wave forecasting system uses the HP-VIT model to forecast regional significant wave height and average zero-crossing period of ocean wave data; The second-generation complete stability prediction system for ships uses a step-by-step correction algorithm to integrate the collected measured ocean environmental parameters into the data set of the HP-VIT model for data fusion. The fused data are then combined with historical environmental data and ship parameter data to input into a pure stability loss sensitivity index prediction model constructed using the regional pure stability loss sensitivity index calculation formula to complete the pure stability loss sensitivity index prediction. At the same time, the fused data are combined with historical environmental data, the obtained ship navigation parameters, and ship parameter data to input into a parameter roll sensitivity index prediction model constructed using the regional parameter roll sensitivity index calculation formula to complete the regional parameter roll prediction.

10. The second generation intact stability prediction system for ships sailing in real sea according to claim 9 is characterized in that: The second-generation complete stability prediction system for ships sailing in real seas is carried on a computer-readable storage medium, and the computer-readable storage medium stores a computer program. When the computer program is executed by a processor, it can realize the functions of the second-generation complete stability prediction system for ships sailing in real seas.

Images