Overturning moment prediction method of vertical ship lift rectangular ship reception chamber-water body-ship coupling system fusing CFD and ITransform-BiLSTM neural network

Abstract

The invention discloses a method for predicting the overturning moment of a rectangular ship reception chamber-water body-ship coupling system of a vertical ship lift by fusing CFD and an ITransform-BiLSTM neural network. The method comprises the following steps: establishing three-dimensional models of different working conditions of the rectangular ship reception chamber-water body-ship system by using three-dimensional modeling software; selecting typical seismic waves as excitation; performing grid independence analysis to obtain a grid division size, and then respectively establishing CFD simulation models under different working conditions according to the grid division size; simulating the most dangerous working condition by utilizing excitation, and inputting CFD simulation models of different working conditions to carry out transient dynamics calculation and upsetting moment calculation; extracting a calculation result, filtering, carrying out standardization processing, and constructing a data set; constructing an overturning moment prediction model based on the neural network, training by using the data set, constructing an overturning moment prediction model based on the neural network, and training by using the data set; and inputting excitation to the trained upsetting moment prediction model, and outputting an upsetting moment prediction result.

Description

Technical Field

[0001] The present invention relates to the intersection of water conservancy engineering and intelligent computing, and in particular to a method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system by integrating CFD and an ITransformer-BiLSTM neural network. Background Art

[0002] As an important navigation facility, vertical ship lifts primarily help ships overcome water level differences in the waterway and navigate smoothly by vertically lifting a vessel-and-water container. However, during operation, the movement of the container, the sloshing of the water within the container, and the sloshing of the vessel interact with each other, forming a complex coupled system of container, water, and ship. Under external excitation, this coupled system generates significant hydrodynamic responses, particularly capsizing moments, which affect structural stability. Because analytical calculations of this system are difficult, current engineering projects generally ignore the impact of the vessel and simplify the analysis to a two-dimensional case of water sloshing within the container. While this simplifies the calculations, it underestimates the hydrodynamic loads and leads to an overestimation of the structural safety factor. Therefore, a more accurate capsizing moment prediction method that considers the three-dimensional rectangular container-water-ship coupling is urgently needed to improve the design reliability and operational safety of ship lifts. Summary of the Invention

[0003] To overcome the above-mentioned deficiencies of the prior art, the present invention provides a method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system by integrating CFD (fluid dynamics calculation) with an ITransformer-BiLSTM neural network.

[0004] According to one aspect of the present invention, the present invention provides a method for predicting the capsizing moment of a vertical ship lift rectangular trolley-water-ship coupling system by integrating CFD and an ITransformer-BiLSTM neural network, comprising: Use 3D modeling software to build 3D models of the rectangular trolley-water-ship system under different working conditions; Select typical earthquake waves as excitation according to the design seismic fortification level of the vertical ship lift and the service area; Select a 3D model, input excitation in the CFD simulation software environment to perform grid independence analysis, obtain the grid size, and then establish CFD simulation models for different working conditions in the CFD simulation software environment based on the grid size; Use excitation to simulate the most dangerous working conditions, input CFD simulation models of different working conditions to perform transient dynamic calculations and overturning moment calculations; Extract calculation results, filter and standardize them, and construct data sets; Construct an overturning moment prediction model based on a neural network and use the dataset for training; Input excitation to the trained overturning moment prediction model and output the overturning moment prediction result.

[0005] As a further technical solution, the three-dimensional model refers to the design tonnage and ship type of the vertical ship lift. Different values ​​within the design tonnage range of the vertical ship lift and different load states of the ship form different working conditions.

[0006] As a further technical solution, CFD simulation model: The fluid volume fraction model is used to track the free surface; The standard k-ε turbulence model is used to describe the turbulent behavior of the fluid; The ship compartment and the ship are considered as rigid bodies and the heat exchange is neglected; Use six-degree-of-freedom solver and user-defined functions to simulate the free swaying of ships; The mesh around the moving ship is updated in real time using dynamic mesh technology, implicit volume update technology is adopted, and mesh smoothing and mesh re-division methods are combined.

[0007] As a further technical solution, the specific steps of grid independence analysis are as follows: In the CFD simulation software environment, the selected three-dimensional model is divided into different grid division numbers to obtain multiple CFD simulation models, and the corresponding grid sizes are recorded; Seismic wave excitation is applied to the longitudinal direction of the rectangular tether, and the overturning moment of each CFD model is calculated. The first four peaks of the overturning moment at the longitudinal bottom center of the tether are extracted as the overturning moment calculation results. The CFD models are compared pairwise according to the overturning moment calculation results in the order of the number of mesh divisions from the smallest to the largest. When the error is less than 1%, it is determined that the increase in the number of mesh divisions has no effect on the overturning moment calculation results. The mesh size corresponding to the one with the lower number of mesh divisions is selected as the mesh size.

[0008] As a further technical solution, the method for simulating the most dangerous working conditions is to apply excitation to the ship compartment in the transverse, longitudinal and vertical directions at the same time. As a further technical solution, the extracted calculation results include the results of transient dynamic calculation and calculation of overturning moment, all in vector form; The results of transient dynamic calculations include: ship length / carrier length, ship width / carrier width, ship draft / liquid level, liquid level / carrier height, earthquake acceleration a(t-1) at time t-1, earthquake acceleration a(t) at time t, acceleration a(t+1) at time t+1, hydrodynamic force F(t-1) at time t-1, and hydrodynamic force F(t) at time t; The results of calculating the overturning moment include: the overturning moment M(t-1) at time t-1, the overturning moment M(t) at time t, and the overturning moment M(t+1) at time t+1.

[0009] As a further technical solution, when training the overturning moment prediction model, the overturning moment M(t+1) at time t+1 in the data set is used as the output of the model, and other calculation result data is used as the input of the model.

[0010] As a further technical solution, the overturning moment prediction model is an ITransformer-BiLSTM neural network model. The network structure consists of an input layer, a position encoding layer, an addition layer, a self-attention layer 1, a discard layer 1, a self-attention layer 2, a discard layer 2, a BiLSTM layer, a fully connected layer and an output layer.

[0011] According to one aspect of the present invention, the present invention provides a capsizing moment prediction system for a vertical ship lift rectangular trolley-water-ship coupling system, comprising: The 3D model building module builds a 3D model of the rectangular ship-carrying compartment-water-ship system based on the designed ship tonnage and ship type of the vertical ship lift; Excitation simulation module, which generates seismic wave type excitation; CFD model building module, to establish the ship type and the CFD simulation model of the rectangular ship-water-ship system; CFD analysis module, used for transient dynamics calculation and overturning moment calculation of CFD simulation models; The data set construction module extracts the calculation results of the CFD analysis module, filters and standardizes them to construct a data set; The overturning moment prediction module builds an overturning moment prediction model based on a neural network, provides overturning moment prediction functions, and also includes: the overturning moment prediction model training function; The result output module outputs the overturning moment prediction results.

[0012] As a further technical solution, the overturning moment prediction module is implemented based on the ITransformer-BiLSTM neural network model. The network structure consists of input layer, position encoding layer, addition layer, self-attention layer 1, discard layer 1, self-attention layer 2, discard layer 2, BiLSTM layer, fully connected layer and output layer.

[0013] Compared with the existing technology, the beneficial effects of the present invention are: the present invention proposes a method for calculating the overturning moment under the coupling action of the rectangular ship support compartment-water body-ship of a vertical ship lift, compiles a calculation program for the overturning moment, and improves the traditional means of calculating the overturning moment from two dimensions to three dimensions, which greatly improves the calculation accuracy, provides a scientific and reasonable basis and reference for the design of vertical ship lifts, and promotes the development of vertical ship lifts; the present invention adopts the ITransformer-BiLSTM model to predict the overturning moment, which has higher prediction accuracy and stronger generalization ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, a brief introduction will be given below to the drawings used in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0015] Figure 1 A flow chart of a method for predicting the capsizing moment of a vertical ship lift rectangular trolley-water-ship coupling system that integrates CFD and an ITransformer-BiLSTM neural network, provided in an embodiment of the present invention; Figure 2 Schematic diagram of the structure of the ITransformer-BiLSTM neural network model in an embodiment of the present invention; Figure 3 Schematic diagram of the ship lift structure and three-dimensional model diagrams of four working conditions in an embodiment of the present invention: (a) 1000t fully loaded, (b) 3000t empty, (c) 3000t fully loaded, and (d) no ship; Figure 4 Schematic diagram showing the comparison of overturning moments of a two-dimensional model and a three-dimensional model under the same earthquake excitation in an embodiment of the present invention; Figure 5 Schematic diagram of a capsizing moment prediction system for a vertical ship lift rectangular support box-water body-ship coupling system in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] It should be noted that: The terms "including" and "having" and any variations thereof in the description and claims of the present invention and the above-mentioned drawings are intended to cover non-exclusive inclusions, for example, a process, method, system, product or apparatus that includes a series of steps or units is not necessarily limited to the steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products or apparatuses.

[0017] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices. The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily need to be executed in the order described. For example, some operations / steps may be further decomposed, while others may be combined or partially combined, so the actual execution order may vary depending on the actual situation.

[0018] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, the technical features in the various embodiments or single embodiments provided by the present invention are arbitrarily combined with each other to form a new technical solution. This combination is not restricted by the sequence of steps and / or structural composition mode, but must be based on the ability of ordinary technicians in this field to implement it. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that this combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0019] Please refer to Figure 1 , Figure 1 The flowchart of a method for predicting the capsizing moment of a vertical ship lift rectangular tether-water-ship coupling system that integrates CFD and an ITransformer-BiLSTM neural network in an embodiment of the present invention includes: Use 3D modeling software to build 3D models of the rectangular trolley-water-ship system under different working conditions; Select typical earthquake waves as excitation according to the design seismic fortification level of the vertical ship lift and the service area; Select a 3D model, input excitation in the CFD simulation software environment to perform grid independence analysis, obtain the grid size, and then establish CFD simulation models for different working conditions in the CFD simulation software environment based on the grid size; Use excitation to simulate the most dangerous working conditions, input CFD simulation models of different working conditions to perform transient dynamic calculations and overturning moment calculations; Extract calculation results, filter and standardize them, and construct data sets; Construct an overturning moment prediction model based on a neural network and use the dataset for training; Input excitation to the trained overturning moment prediction model and output the overturning moment prediction result.

[0020] Specifically, the three-dimensional model refers to the designed ship passing tonnage and ship type design of the vertical ship lift. Different values ​​within the designed ship passing tonnage range of the vertical ship lift and different load states of the ship form different working conditions.

[0021] Specifically, the CFD simulation model: The fluid volume fraction model is used to track the free surface; The standard k-ε turbulence model is used to describe the turbulent behavior of the fluid; The ship compartment and the ship are considered as rigid bodies and the heat exchange is neglected; Use six-degree-of-freedom solver and user-defined functions to simulate the free swaying of ships; The mesh around the moving ship is updated in real time using dynamic mesh technology, implicit volume update technology is adopted, and mesh smoothing and mesh re-division methods are combined.

[0022] Specifically, the steps of grid independence analysis are as follows: In the CFD simulation software environment, the selected three-dimensional model is divided into different grid division numbers to obtain multiple CFD simulation models, and the corresponding grid sizes are recorded; Seismic wave excitation is applied to the longitudinal direction of the rectangular tether, and the overturning moment of each CFD model is calculated. The first four peaks of the overturning moment at the longitudinal bottom center of the tether are extracted as the overturning moment calculation results. The CFD models are compared pairwise according to the overturning moment calculation results in the order of the number of mesh divisions from the smallest to the largest. When the error is less than 1%, it is determined that the increase in the number of mesh divisions has no effect on the overturning moment calculation results. The mesh size corresponding to the one with the lower number of mesh divisions is selected as the mesh size.

[0023] Specifically, the method for simulating the most dangerous working condition is: applying excitation to the carrier compartment in the transverse, longitudinal and vertical directions simultaneously.

[0024] As a further technical solution, the extracted calculation results include the results of transient dynamic calculation and calculation of overturning moment, all in vector form; The results of transient dynamic calculations include: ship length / carrier length, ship width / carrier width, ship draft / liquid level, liquid level / carrier height, earthquake acceleration a(t-1) at time t-1, earthquake acceleration a(t) at time t, acceleration a(t+1) at time t+1, hydrodynamic force F(t-1) at time t-1, and hydrodynamic force F(t) at time t; The results of calculating the overturning moment include: the overturning moment M(t-1) at time t-1, the overturning moment M(t) at time t, and the overturning moment M(t+1) at time t+1.

[0025] Specifically, when training the overturning moment prediction model, the overturning moment M(t+1) at time t+1 in the data set is used as the output of the model, and other calculation result data are used as the input of the model.

[0026] Specifically, the overturning moment prediction model is an ITransformer-BiLSTM neural network model, such as Figure 2 As shown, Figure 2 This is a schematic diagram of the structure of the Transformer-BiLSTM neural network model in an embodiment of the present invention. The Transformer portion of the neural network includes two self-attention mechanisms and a dropout layer, which are connected in series with the BiLSTM structure to form a hybrid model. The network structure consists of an input layer, a position encoding layer, an addition layer, self-attention layer 1, dropout layer 1, self-attention layer 2, dropout layer 2, a BiLSTM layer, a fully connected layer, and an output layer.

[0027] A method for predicting the overturning moment of a vertical ship lift rectangular ship support box-water body-ship coupling system comprises the following steps: (1) The ship-carrying compartment, water body inside the compartment and ship in the national key R&D program "200m-class large vertical ship lift complete set technology" are taken as the research objects. According to the design requirements, the tonnage of the ship is between 1000t and 3000t, and four working conditions are set: no ship empty, 1000t full load, 3000t empty and 3000t full load. Use SolidWorks to build three-dimensional models of different working conditions. When modeling, the ship-carrying compartment and ship are simplified into rectangles. The ship size and draft refer to the design requirements as follows: Figure 3 As shown, Figure 3 Schematic diagram of the ship lift structure and three-dimensional model diagrams of four working conditions in an embodiment of the present invention: (a) 1000t full load, (b) 3000t no load, (c) 3000t full load, and (d) no ship.

[0028] (2) Considering the location of the vertical ship lift and the seismic fortification level required by the design, a typical 10-second seismic wave that meets the regional characteristics of the service area is selected as the excitation and saved in the PROFILE file format that can be read by ANSYS-Fluent.

[0029] (3) Ansys-Fluent was selected as the CFD simulation software: the volume fraction (VOF) model was used to track the free surface. The standard k-ε turbulence model was used to describe the turbulent behavior of the fluid. The ship compartment and the ship were treated as rigid bodies and heat exchange was ignored. The free swaying of the ship was simulated using a six-degree-of-freedom (6-DOF) solver and user-defined functions (UDFs). The mesh around the moving ship was updated in real time using dynamic mesh technology. Mesh smoothing and mesh re-division methods were combined to prevent the generation of negative volume in the mesh. Implicit volume update technology was used to improve the convergence and stability of transient calculations.

[0030] (4) Randomly select one of the different three-dimensional ship-carrier-water-ship models established, divide them into grids of different sizes, and thus generate calculation models with different numbers of grids. Then, perform grid independence analysis, apply typical seismic wave excitation, calculate each model, and extract the first four peaks of the overturning moment at the longitudinal bottom center of the ship-carrier as comparison indicators. It is required that as the number of grids increases, when the error of the calculation result is less than 1%, the grid size with the smaller number of grids is selected as the basis for the final grid division. Taking the 3000t no-load model as an example, the number of grids is set to 504808, 650728, 1120848 and 3443384, respectively, and the corresponding grid sizes are 0.36m, 0.32m, 0.27m and 0.18m, respectively. A 7-magnitude EL-Centro seismic wave is applied to the longitudinal direction of the ship-carrier, with a peak acceleration amplitude of 0.1g. After calculation, the first four peaks of the overturning moment at the longitudinal bottom center of the ship-carrier are extracted as comparison indicators. As shown in Table 1, Table 1 is the peak value of the longitudinal overturning moment at the bottom center of the 3000t empty ship compartment and the peak value comparison table under different grid numbers.

[0031] Table 1 Comparison of the peak longitudinal overturning moment at the bottom center of a 3000t empty ship compartment under different grid numbers and peak values

[0032] As shown in Table 1, as the number of grid cells increases, the errors between the four peak overturning moment calculation results gradually decrease, and the results tend to be stable. In particular, when the number of grid cells is increased from 1120848 to 3443384, the error between the two results is less than 1%. It can be considered that the increase in the number of grid cells at this point has little effect on the results. To balance calculation accuracy and efficiency, this paper selects a number of cells of approximately 1120848, which corresponds to a grid size of approximately 0.27m. Based on this grid size, a CFD simulation model of a three-dimensional rectangular tether-water-ship system under different working conditions is established.

[0033] When the ship compartment is subjected to lateral, longitudinal and vertical earthquake excitations at the same time, the hydrodynamic load caused by the coupled motion of the water body and the ship in the compartment is the largest. This paper selects this working condition as the most dangerous working condition, adopts transient dynamic calculation to ensure the safety of the design, and then calculates the overturning moment of the four models under the most dangerous working condition. For the 3000t no-load model, the calculation formula derived under the previous two-dimensional model is compared and analyzed with the results of the CFD simulation of the three-dimensional model of the present invention. Figure 4 shown.

[0034] The results show that the CFD simulation results of the three-dimensional model proposed in this invention are in good agreement with the calculation results of the previous two-dimensional model in terms of trend, but the calculation results of the three-dimensional model are significantly higher than those of the two-dimensional model, especially in the calculation of the longitudinal overturning moment, the calculation peak of the three-dimensional model is about 20% to 30% higher. This is because the three-dimensional model takes into account the coupled sloshing effect of the ship and the water body, and can more realistically reflect the actual hydrodynamic load characteristics. In contrast, the two-dimensional model ignores the impact of the ship on the sloshing of the water body, resulting in a lower prediction value. This simplified assumption may lead to an underestimation of the working load on the ship compartment in actual engineering design, thereby affecting the safety of the design.

[0035] (5) Extracting calculation data. For the calculation data, the Z-score method is first used to detect outliers to ensure the accuracy and reliability of the data. Then the data is standardized to remove the dimension effect. The 11 vectors including ship length / carrier length, ship width / carrier width, ship draft / liquid level, liquid level / carrier height, accelerations a(t-1) and a(t) at time t-1 and time t, hydrodynamic forces F(t-1) and F(t), capsizing moment M(t-1) and M(t), and acceleration a(t+1) at time t+1 are used as the input of the model. The capsizing moment M(t+1) at time t+1 is used as the single output of the model to construct a data set for training the prediction model. In order to ensure the generalization ability and reliability of the neural network model, the data set is divided into training set, validation set and test set, and distributed in a ratio of 7:2:1.

[0036] (6) In the parameter selection of the ITransformer-BiLSTM model, the self-attention mechanism was adopted, with 2 attention heads and 4 keys. The number of neurons in the input layer was equal to the number of input vectors, the hidden layer was set to 64 neurons, and the output layer of both neural network models was set to 1 neuron. The maximum number of training generations was 70, the batch size was set to the default value of 50, and the target value of the training error was set to 1×10 -5 .

[0037] (7) For the dataset of predicted overturning moment, LSTM, BiLSTM, ELMAN and ITransformer-BiLSTM models were established for comparison. The normalized evaluation indicators of the overturning moment prediction effect are shown in Table 2.

[0038] Table 2 Five evaluation indicators of overturning moment prediction performance under different algorithms

[0039] As shown in the table, on the training set, the IT-B model achieved a mean absolute error (MAE) of 0.14180, a mean absolute percentage error (MAPE) of 0.64664%, a mean square error (MSE) of 0.04504, a root mean square error (RMSE) of 0.21222, and a coefficient of determination (R²) of 0.97044, all significantly outperforming the other three models. Results on the validation and test sets also show that the IT-B model maintains the best performance in terms of both error metrics and goodness of fit. The BiLSTM model outperformed the LSTM and ELMAN models, but still lagged behind the IT-B model. The LSTM and ELMAN models performed relatively poorly. The ELMAN model, in particular, achieved a MAPE of 4.38460% on the validation set and an R² of only 0.75928 on the test set, indicating weak generalization ability in the overturning moment prediction task.

[0040] (8) Save the training results of the ITransformer-BiLSTM neural network based on MATLAB in .m format so that it can be called when calculating the overturning moment under other excitations.

[0041] The implementation basis of each embodiment of the present invention is achieved through programmed processing by a device with processor functionality. Therefore, in engineering practice, the technical solutions and functions of each embodiment of the present invention are encapsulated into various modules. Based on this reality, on the basis of the above embodiments, an embodiment of the present invention provides a capsizing moment prediction system for a ship lift rectangular support box-water body-ship coupling system. The system is used to execute a capsizing moment prediction method for a vertical ship lift rectangular support box-water body-ship coupling system that integrates CFD and ITransformer-BiLSTM neural network in the above method embodiment. The system includes: The 3D model building module builds a 3D model of the rectangular ship-carrying compartment-water-ship system based on the designed ship tonnage and ship type of the vertical ship lift; Excitation simulation module, which generates seismic wave type excitation; CFD model building module, to establish the ship type and the CFD simulation model of the rectangular ship-water-ship system; CFD analysis module, used for transient dynamics calculation and overturning moment calculation of CFD simulation models; The data set construction module extracts the calculation results of the CFD analysis module, filters and standardizes them to construct a data set; The overturning moment prediction module builds an overturning moment prediction model based on a neural network, provides overturning moment prediction functions, and also includes: the overturning moment prediction model training function; The result output module outputs the overturning moment prediction results.

[0042] It should be noted that the system embodiments provided by the present invention are not only used to implement the methods in the above-mentioned method embodiments, but are also used to implement the methods in other method embodiments provided by the present invention. The only difference is the setting of corresponding functional modules. The principles thereof are basically the same as those of the above-mentioned system embodiments provided by the present invention. As long as those skilled in the art refer to the specific technical solutions in other method embodiments on the basis of the above-mentioned system embodiments, obtain corresponding technical means and technical solutions composed of these technical means by combining technical features, and ensure the practicality of the technical solutions, they can improve the equipment in the above-mentioned system embodiments to obtain corresponding equipment class embodiments for implementing the methods in other method class embodiments. For example: The overturning moment prediction module is implemented based on the ITransformer-BiLSTM neural network model. The network structure consists of input layer, position encoding layer, addition layer, self-attention layer 1, discard layer 1, self-attention layer 2, discard layer 2, BiLSTM layer, fully connected layer and output layer.

[0043] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, located in one place or distributed across multiple network units. Depending on practical needs, some or all of these modules may be selected to achieve the objectives of this embodiment. Persons of ordinary skill in the art will understand and implement these embodiments without inventive effort.

[0044] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0045] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0046] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0047] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system by integrating CFD and ITransformer-BiLSTM neural network, characterized by: include: Use 3D modeling software to build 3D models of the rectangular trolley-water-ship system under different working conditions; Select typical earthquake waves as excitation according to the design seismic fortification level of the vertical ship lift and the service area; Select a 3D model, input excitation in the CFD simulation software environment to perform grid independence analysis, obtain the grid size, and then establish CFD simulation models for different working conditions in the CFD simulation software environment based on the grid size; Use excitation to simulate the most dangerous working conditions, input CFD simulation models of different working conditions to perform transient dynamic calculations and overturning moment calculations; Extract calculation results, filter and standardize them, and construct data sets; Construct an overturning moment prediction model based on a neural network and use the dataset for training; Input excitation to the trained overturning moment prediction model and output the overturning moment prediction result.

2. A method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 1, characterized in that: The three-dimensional model is designed with reference to the designed ship passing tonnage and ship type of the vertical ship lift. Different values ​​within the designed ship passing tonnage range of the vertical ship lift and different load states of the ship form different working conditions.

3. The method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 1, characterized in that: The CFD simulation model: The fluid volume fraction model is used to track the free surface; The standard k-ε turbulence model is used to describe the turbulent behavior of the fluid; The ship compartment and the ship are considered as rigid bodies and the heat exchange is neglected; Use six-degree-of-freedom solver and user-defined functions to simulate the free swaying of ships; The mesh around the moving ship is updated in real time using dynamic mesh technology, implicit volume update technology is adopted, and mesh smoothing and mesh re-division methods are combined.

4. The method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 3, characterized in that: The specific steps of grid independence analysis are: In the CFD simulation software environment, the selected three-dimensional model is divided into different grid division numbers to obtain multiple CFD simulation models, and the corresponding grid sizes are recorded; Seismic wave excitation is applied to the longitudinal direction of the rectangular tether, and the overturning moment of each CFD model is calculated. The first four peaks of the overturning moment at the longitudinal bottom center of the tether are extracted as the overturning moment calculation results. The CFD models are compared pairwise according to the overturning moment calculation results in the order of the number of mesh divisions from the smallest to the largest. When the error is less than the preset value, it is determined that the increase in the number of mesh divisions has no effect on the overturning moment calculation results. The mesh size corresponding to the one with the lower number of mesh divisions is selected as the mesh size.

5. The method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 1, characterized in that: The method for simulating the most dangerous working condition is specifically: applying excitation to the carrier compartment in the transverse, longitudinal and vertical directions simultaneously.

6. The method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 1, characterized in that: The extracted calculation results include the results of transient dynamic calculation and overturning moment calculation, both in vector form; The calculation results of transient dynamics calculation include: ship length / carrier length, ship width / carrier width, ship draft / liquid level, liquid level / carrier height, t -1 moment earthquake acceleration a ( t -1) t Earthquake acceleration at time a ( t ), t Acceleration at time +1 a ( t +1), t -1 moment of water power F ( t -1) and t The water power of the moment F ( t ); The results of the overturning moment calculation include: t -1 moment overturning moment M ( t -1) t Overturning moment M ( t )as well as t Overturning moment at time +1 M ( t +1).

7. A method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 6, characterized in that: When training the overturning moment prediction model, the data set t Overturning moment at time +1 M ( t +1) is used as the output of the model, and other calculation result data is used as the input of the model.

8. The method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 1, characterized in that: The overturning moment prediction model is an ITransformer-BiLSTM neural network model, and the network structure consists of an input layer, a position encoding layer, an addition layer, a self-attention layer 1, a discard layer 1, a self-attention layer 2, a discard layer 2, a BiLSTM layer, a fully connected layer and an output layer.

9. A capsizing moment prediction system for a vertical ship lift rectangular ship support-water-ship coupling system, characterized in that: A method for predicting the overturning moment of a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and an ITransformer-BiLSTM neural network as described in any one of claims 1 to 8, comprising: The 3D model building module builds a 3D model of the rectangular ship-carrying compartment-water-ship system based on the designed ship tonnage and ship type of the vertical ship lift; Excitation simulation module, which generates seismic wave type excitation; CFD model building module, to establish the ship type and the CFD simulation model of the rectangular ship-water-ship system; CFD analysis module, used for transient dynamics calculation and overturning moment calculation of CFD simulation models; The data set construction module extracts the calculation results of the CFD analysis module, filters and standardizes them to construct a data set; The overturning moment prediction module builds an overturning moment prediction model based on a neural network, provides overturning moment prediction functions, and also includes: the overturning moment prediction model training function; The result output module outputs the overturning moment prediction results.

10. The capsizing moment prediction system for a vertical ship lift rectangular trolley-water-ship coupling system integrating CFD and a Transformer-BiLSTM neural network as claimed in claim 9, characterized in that: The overturning moment prediction module is implemented based on the ITransformer-BiLSTM neural network model. The network structure consists of an input layer, a position encoding layer, an addition layer, a self-attention layer 1, a discard layer 1, a self-attention layer 2, a discard layer 2, a BiLSTM layer, a fully connected layer and an output layer.

Images