Full coupling simulation system and method for abnormal wave and sloshing wave of liquid cargo ship based on viscous flow theory and medium

Abstract

The application discloses a kind of based on viscous flow theory abnormal wave and sloshing wave full coupling simulation system, method and medium for liquid cargo ship, belong to ship and ocean engineering extreme wave evolution simulation technical field, including: extreme wave pool module, including wave generation area, coupling solution area and wave absorbing area in turn along wave propagation direction;Ship-hold-tank-sloshing plate model module;Three areas of external fluid, overlapping grid and internal fluid are established and VOF is used to capture free surface;Abnormal wave generation module generates abnormal wave of same target parameter using linear physical mechanism and nonlinear modulation mechanism respectively;Two-way full coupling solution module, according to external wave load and internal sloshing feedback load, real-time updates ship motion state;KAN proxy model optimization module, continuously predicts and optimizes sloshing plate parameters and outputs optimization scheme after checking. The application realizes abnormal wave-ship-hold-tank three-way synchronous two-way full coupling simulation and intelligent optimization closed loop of sloshing plate parameter.

Description

Technical Field

[0001] This invention relates to a fully coupled simulation system, method, and medium for anomalous waves and sloshing waves in liquid cargo ships based on viscous flow theory, belonging to the field of extreme wave evolution simulation technology in shipbuilding and marine engineering. Background Technology

[0002] When liquid cargo ships, LNG carriers, and other liquid-carrying vessels navigate in extreme sea states, they may encounter anomalous waves characterized by unusual wave height, concentrated energy, and short duration. These anomalous waves, acting on the hull, generate significant transient wave loads within a short period, inducing hull roll, heave, and sway responses. For vessels with internal liquid tanks, hull motion further exacerbates liquid sloshing within the tanks, leading to phenomena such as liquid level rise, liquid impact, increased bulkhead pressure, and drastic changes in the internal flow field. The pressure, inertial forces, and torques generated by the sloshing liquid within the tanks then feed back into the hull's motion, creating a complex two-way coupling relationship where external anomalous waves and liquid tank sloshing waves act on the hull and influence each other.

[0003] In existing technologies, anomalous wave generation, ship motion calculation, and tank sloshing analysis typically employ relatively independent modeling methods. Some methods focus only on external wave generation and propagation, failing to incorporate tank sloshing wave feedback into the ship motion solution; others use a given form of motion excitation as the tank sloshing input, making it difficult to reflect the true driving process of external anomalous waves on ship motion; still others simplify the tank liquid to a fixed or added mass, failing to capture physical phenomena such as free surface sloshing, surface rise, localized impacts, and flow around sloshing plates. Although some coupling techniques have emerged in recent years, they have all undergone simplification, and a fully coupled simulation analysis technique based on viscous flow theory for strongly nonlinear anomalous waves and sloshing waves is still lacking.

[0004] Regarding the optimization of sloshing baffles in liquid tanks, Chinese patent application publication number CN117235935A discloses a method for constructing a floating liquid tank sloshing baffle model and optimizing its dimensions. This technology constructs a numerical model of liquid tank sloshing based on viscous flow theory, uses the volumetric fluid flow (VOF) method to capture the free surface, and simulates the motion of the sloshing baffle through an overlapping mesh method. However, it has the following shortcomings: First, its external excitation uses sinusoidal regular waves instead of anomalous waves with concentrated energy characteristics, which cannot simulate the response of liquid cargo ships under extreme sea conditions of anomalous waves; second, it only focuses on the sloshing of the liquid tank itself, without involving the hull model and the coupling effect between the hull and external waves, and cannot reflect the two-way feedback relationship between the hull motion and the sloshing of the liquid tank; third, its sloshing baffle optimization uses ergonomic methods and single-path optimization methods, which have low computational efficiency and lack a rapid prediction mechanism using a surrogate model.

[0005] Regarding the integration of neural networks and CFD, Chinese patent application publication number CN114818128A discloses a neural network modeling method and optimization method for optimizing local curved surfaces of ship hulls. This technology uses a BP neural network combined with a Kriging model to establish a surrogate model, taking the coordinates of the ship's control points as input and hydrodynamic characteristics as output to optimize the hull surface. However, the optimization object is the hull shape, not the parameters of the liquid tank oscillation plates; it uses a combination of BP neural networks and Kriging models instead of a KAN network; and it lacks a closed-loop mechanism to re-input the surrogate model's prediction results into a fully coupled CFD model for verification.

[0006] In the area of ​​CFD simulation of hull-tank coupling, Chinese patent application publication number CN120611560A discloses a CFD-based method for calculating fluid sloshing loads in C-type tanks of LNG ships. This technology, based on the STAR-CCM+ simulation platform, employs the VOF method to capture the free surface, overlapping meshes to handle hull motion, and a DFBI model to achieve fluid-structure interaction, analyzing the coupling effect between hull motion and multi-tank fluid sloshing when a ship navigates in waves approaching real sea conditions. However, it has the following shortcomings: First, the wave conditions are random or regular waves approaching real sea conditions, rather than anomalous waves with specific focusing characteristics, and it does not address the generation and control of anomalous waves; second, it only analyzes sloshing loads and does not involve the sloshing plate structure and its optimization; third, it does not establish a surrogate model for rapid parameter prediction and automatic optimization.

[0007] In summary, the existing technology lacks an integrated simulation method that can realize the dual-mechanism generation of anomalous waves, the control of the ship's motion degrees of freedom, the bidirectional fully coupled solution of anomalous waves, ship, and liquid tanks, and the intelligent optimization closed loop of the oscillation plate parameters based on the KAN surrogate model, all within the framework of a unified viscous flow theory. Summary of the Invention

[0008] The present invention aims to solve the technical problems in the prior art where the generation of abnormal waves, the calculation of ship motion and the analysis of liquid tank sloshing are independent and cannot achieve synchronous two-way fully coupled simulation of the three, as well as the technical problems in the prior art where the optimization efficiency of the sloshing plate parameters is low and there is a lack of automatic optimization closed loop.

[0009] To achieve the above objectives, the present invention provides the following technical solution: The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory described in this invention includes: An extreme wave pool module is configured to generate and propagate abnormal waves. The extreme wave pool includes a wave-generating zone, a coupled solution zone, and a wave-dissipating zone in sequence along the wave propagation direction. The hull-liquid tank-squeezing plate model module is located within the coupled solution area and includes a hull model, a liquid tank model located inside the hull, and a squeezing plate model located inside the liquid tank. The fluid region partitioning module is configured to create an external fluid region, an overlapping grid region, and an internal fluid region. Both the external and internal fluid regions use the fluid volume method to capture the free liquid surface. The abnormal wave generation module is configured to generate abnormal waves with the same target parameters using linear physical mechanisms and nonlinear modulation mechanisms respectively, and input the generated initial wave train into the extreme wave pool for evolution until the target waveform requirements are met. The bidirectional fully coupled solution module is configured to generate and propagate abnormal waves in the external fluid region driven by the wave-generating boundary at each time step. The hull motion state is updated in real time according to the external wave load and the internal sloshing feedback load of the liquid tank. After the hull motion state is updated, the overlapping grid region, the liquid tank wall and the sloshing plate boundary are updated synchronously, realizing the bidirectional fully coupled solution between the abnormal wave evolution excitation, the hull motion response and the liquid tank sloshing wave feedback. The KAN surrogate model optimization module is configured to construct a KAN surrogate model. It establishes a nonlinear mapping relationship with abnormal wave parameters, liquid tank parameters, and oscillation plate parameters as inputs and a fully coupled response index as output. The KAN surrogate model is used to continuously predict and optimize the oscillation plate parameters. The candidate optimization schemes are re-input into the bidirectional fully coupled solution module for verification, and the optimized oscillation plate arrangement scheme that meets the requirements is output.

[0010] Furthermore, in the anomalous wave generation module, the linear physical mechanism for generating anomalous waves includes at least one of the new wave theory method and the transient wave packet method. The new wave theory method discretizes the target wave spectrum into multiple frequency component waves, determines the frequency, wave number, amplitude, and initial phase of each component wave, and uses phase adjustment technology to superimpose each component wave at a specified position and time to form an anomalous wave. The transient wave packet method constructs an initial wave train based on a frequency sweep signal that varies with time, so that different frequency components gradually converge to form an anomalous wave during propagation. The nonlinear modulation mechanism for generating anomalous waves is based on nonlinear wave theory or nonlinear envelope wave method, and controls the evolution of the carrier and wave packet envelope under nonlinear and dispersion effects to concentrate wave energy at the target position to form an anomalous wave.

[0011] Furthermore, in the fluid region partitioning module, the external fluid region is configured to calculate the propagation, evolution, focusing, diffraction, radiation, and hydrodynamic effects of anomalous waves on the hull; the overlapping grid region is configured to surround the hull, liquid tanks, and sway dampers and move synchronously with the hull movement, used to handle real-time grid updates during hull movement; the internal fluid region is configured to calculate free surface sloshing, surface rise, liquid impact, bulkhead pressure, and vortex discharge near the sway dampers inside the liquid tanks; both the external and internal fluid regions are configured with a gas and liquid dual-continuum two-phase model.

[0012] Furthermore, it also includes a coordinate system and motion control module, configured to establish a reference coordinate system fixed in the extreme wave pool and a ship-mounted coordinate system fixed in the hull, and set dynamic fluid-structure interaction motion solution control conditions; the hull motion degrees of freedom include sway, roll, heave, pitch, pitch and yaw, and by setting the activation state of the degrees of freedom, single-degree-of-freedom, double-degree-of-freedom and multi-degree-of-freedom working conditions are formed to realize the separation and coupling calculation of each motion degree of freedom.

[0013] Furthermore, in the bidirectional fully coupled solution module, the information transmission relationship between the abnormal wave evolution excitation, the hull motion response, and the feedback effect of the liquid tank sloshing wave is expressed as follows: ; ; in, Let the generalized motion vector of the ship's hull be . and These represent the external flow field and the internal flow field states, respectively. For the generalized excitation load generated by external anomalous waves, The generalized feedback load generated by the sloshing of the liquid tank. This represents the process of solving for the ship's motion. This represents the process of updating the moving boundary and the flow field. This represents the number of coupling iterations within the same physical time step.

[0014] Furthermore, in the KAN proxy model optimization module, the basic mapping relationship of the KAN proxy model is expressed as follows: ; in, For network layer number, and The first Layer and first Layer node number, This is the output of the i-th node in the l-th layer. The number of nodes in the l-th layer. To connect the first Layer The node and the first Layer The KAN surrogate model has learnable univariate functions for each node; the input parameters of the KAN surrogate model include at least the period of the abnormal spectral peak, the liquid tank filling rate, and the relative height of the oscillation plate, and the output parameters include at least the roll amplitude, heave amplitude, sway amplitude, and liquid level rise amplitude; the network node structure of the KAN surrogate model is set to 3-16-12-4, that is, the input layer has 3 nodes, the first intermediate layer has 16 nodes, the second intermediate layer has 12 nodes, and the output layer has 4 nodes; the learnable univariate functions on each connection edge are parameterized using third-order B-spline functions.

[0015] Furthermore, in the KAN surrogate model optimization module, when continuously predicting and optimizing the oscillator parameters using the KAN surrogate model, a comprehensive evaluation function is used to evaluate the candidate oscillator schemes: ; in, and These represent the hull roll amplitude under the conditions of controlled sway plates and the baseline condition of uncontrolled sway plates, respectively. and These represent the hull heave amplitudes under controlled and uncontrolled baseline conditions, respectively. and These represent the hull roll amplitudes under controlled and uncontrolled baseline conditions, respectively. and These represent the liquid level rise amplitude on the tank wall under the conditions of controlled sloshing plates and the baseline condition of uncontrolled sloshing plates, respectively; comprehensive evaluation function. The smaller the value, the better the overall suppression effect of the sloshing plate scheme on the coupled response of hull motion and liquid tank sloshing.

[0016] Furthermore, in the KAN surrogate model optimization module, when the candidate optimization scheme is re-input into the fully coupled viscous flow model for verification, if the relative error between the predicted value of the KAN surrogate model and the fully coupled calculated value meets the preset requirements, the corresponding damping plate height, position and arrangement are output; if the error does not meet the requirements, the new fully coupled calculation results are added to the sample database and the KAN surrogate model is retrained.

[0017] This invention also provides a fully coupled simulation method for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory, applied to the aforementioned fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory, comprising: S1: Establish an extreme wave pool model, which includes a wave-generating zone, a coupled solution zone, and a wave-dissipating zone along the wave propagation direction. S2: Establish a hull model, a liquid tank model, and a sway plate model within the coupled solution region, and establish an external fluid region, an overlapping mesh region, and an internal fluid region. The external fluid region and the internal fluid region both use the fluid volume method to capture the free liquid surface. S3: Establish a reference coordinate system fixed in the extreme wave pool and a ship-mounted coordinate system fixed in the hull, and set dynamic fluid-structure interaction motion solution control conditions. The hull motion degrees of freedom include pitching, swaying, heaving, rolling, pitching and bowing. By setting the activation state of the degrees of freedom, single-degree-of-freedom, double-degree-of-freedom and multi-degree-of-freedom working conditions are formed to realize the separation and coupling calculation of each motion degree of freedom. S4: Abnormal waves with the same target parameters are generated using linear physical mechanisms and nonlinear modulation mechanisms respectively, and the generated initial wave train is input into the extreme wave pool for evolution until the target waveform requirements are met; S5: In each time step, the external fluid region is driven by the wave-generating boundary to generate and propagate abnormal waves. The hull motion state is updated in real time according to the external wave load and the internal sloshing feedback load of the liquid tank. After the hull motion state is updated, the overlapping grid region, the liquid tank wall and the sloshing plate boundary are updated synchronously to realize the bidirectional full coupling solution between the abnormal wave evolution excitation, the hull motion response and the liquid tank sloshing wave feedback. S6: Construct a KAN surrogate model, using anomalous wave parameters, liquid tank parameters, and oscillation plate parameters as inputs and the fully coupled response index as output to establish a nonlinear mapping relationship. Use the KAN surrogate model to continuously predict and optimize the oscillation plate parameters. Re-input the candidate optimization scheme into the fully coupled viscous flow model for verification and output an optimized oscillation plate arrangement scheme that meets the requirements.

[0018] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described fully coupled simulation method for anomalous waves and sloshing waves for liquid cargo ships based on viscous flow theory.

[0019] Compared with existing technologies, the fully coupled simulation system, method, and medium for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory of the present invention exhibit the following beneficial effects in terms of technical performance and practical application: First, this invention develops a numerical water tank generation technique for extreme anomaly waves based on viscous flow theory and oriented towards specified wave parameters. This technique can generate target anomaly waves according to given wave height, period, focusing position, and focusing time, and verify the wave generation accuracy by comparing the target waveform with the waveform monitored in the extreme water tank. Simultaneously, this invention further develops extreme anomaly wave generation methods based on different physical mechanisms. Both linear physical mechanisms (including new wave theory and transient wave packet technology) and nonlinear modulation mechanisms can be used to generate anomaly waves with the same target parameters, thus making the anomaly waves generated by different mechanisms comparable as external excitations to the ship in terms of wave height, period, and focusing conditions, avoiding differences in coupling response caused by inconsistent excitation wave parameters. Compared to existing technologies that only use sinusoidal regular wave excitation, this invention can more realistically simulate extreme sea states with anomaly waves.

[0020] Secondly, this invention develops a fully coupled simulation technology for anomalous waves and tank sloshing waves in liquid cargo ships. This technology can simultaneously solve for the propagation of external anomalous waves, the multi-degree-of-freedom motion of the hull, and the evolution of sloshing waves inside the tanks within a unified viscous flow theory calculation framework. It also feeds back the pressure, inertial force, and torque generated by the tank sloshing to the hull's motion equations, thereby establishing a two-way fully coupled relationship between external wave excitation, hull motion response, and tank sloshing loads. Simultaneously, this invention establishes a technology for separating and coupling the hull's degrees of freedom of motion, enabling separate calculations of single-degree-of-freedom, two-degree-of-freedom, and multi-degree-of-freedom motion constraints. This allows for the extraction of the independent contributions and coupling effects of roll, heave, and sway degrees of freedom on the tank sloshing response and the amplification of unfavorable hull motion, achieving a more comprehensive coupled simulation capability.

[0021] Third, this invention develops an intelligent optimization technology for liquid tank oscillation plates based on the KAN surrogate model. This technology uses parameters such as the period of abnormal spectral peaks, liquid tank filling rate, and relative height of the oscillation plate as inputs, and the fully coupled responses of hull roll, heave, sway, and liquid level rise as outputs, establishing a nonlinear mapping relationship between design parameters and response indicators. The KAN surrogate model is used to rapidly predict the continuous parameter space of the oscillation plate, and candidate optimization parameters are determined by combining this with a comprehensive evaluation function. The candidate schemes are then re-input into the fully coupled viscous flow model for high-precision verification, forming an intelligent optimization closed loop, thereby reducing a large number of repetitive CFD calculations. Compared to the ergonomic method of existing technologies, the optimization efficiency of this invention is significantly improved; compared to the BPNN-Kriging model of existing technologies, this invention uses a KAN network and introduces a fully coupled verification closed-loop mechanism, resulting in higher prediction accuracy and reliability. Attached Figure Description

[0022] Figure 1 This is a simplified flowchart of the fully coupled simulation method of anomalous waves and sloshing waves for liquid cargo ships based on viscous flow theory in an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of the information transmission relationship of the fully coupled solution module in an embodiment of the present invention.

[0024] Figure 3 This is a flowchart of the intelligent optimization process for the oscillator based on the KAN proxy model in an embodiment of the present invention.

[0025] Figure 4 This is a schematic diagram of the spatial physical layout of the extreme wave pool model in an embodiment of the present invention.

[0026] Figure 5 This is a cross-sectional structural dimension diagram of the hull and liquid tank in an embodiment of the present invention.

[0027] Figure 6 This is a schematic diagram of the grid setting and division of key areas in an embodiment of the present invention.

[0028] Figure 7 This is a comparison diagram of the effects of abnormal waveforms in the same time domain generated by three methods in the embodiments of the present invention.

[0029] Figure 8 This is a time history curve of the ship's roll angle under the action of an abnormal wave in an embodiment of the present invention.

[0030] Figure 9 This is a time history curve of the heave displacement of the ship under the action of an abnormal wave in an embodiment of the present invention.

[0031] Figure 10 This is a time history curve of the hull sway displacement under the action of an abnormal wave in an embodiment of the present invention.

[0032] Figure 11 This is a time history curve of the liquid level rise height on the tank wall under the action of an abnormal wave in an embodiment of the present invention.

[0033] Figure 12 This is a time history curve of swaying torque under the action of abnormal waves in an embodiment of the present invention.

[0034] Figure 13 This is a spatial velocity vector coupling diagram of the external and internal flow fields at a certain moment in an embodiment of the present invention. Detailed Implementation

[0035] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0036] Example 1: This embodiment provides a fully coupled simulation method for anomalous waves and sloshing waves in liquid cargo ships based on viscous flow theory. Figure 1 The diagram shown is a simplified overall flowchart of the method of the present invention, which includes the following steps S1 to S6.

[0037] Step S1, establish an extreme wave pool model: like Figure 4 As shown in the figure, the spatial physical layout of the extreme wave tank model is illustrated. Along the wave propagation direction, three functional sections are labeled: the wave-generating zone, the coupled solution zone, and the wave-dissipating zone. Within the wave-generating zone, wave-generating plates and their directional arrows indicate the wave generation method. The coupled solution zone indicates the positions of the hull model and the liquid tank model. The wave-dissipating zone indicates the end-stage wave-dissipating function. This step designs the boundary dimensions based on the test requirements and establishes the initial model of the extreme wave tank. Based on the size of the test object, the wave height, period, generation location, generation time, evolution distance, and wave-dissipating requirements of the extreme anomalous waves, reasonable boundary dimensions are designed to establish a three-dimensional extreme wave numerical tank. The extreme wave tank, along the wave propagation direction, includes the wave-generating zone, the coupled solution zone, and the wave-dissipating zone. The wave-generating zone is used to automatically generate the initial wave train of the target anomalous wave through the parameter adjustment module combined with the extreme wave database, and to perform wave nonlinear evolution; the coupled solution zone is used to arrange the hull model, liquid tank model and sway-damping plate model, and to carry out fully coupled calculations of the three physical fields considering the viscosity effect; the wave-dissipating zone is used to weaken the influence of the terminal reflected wave on the hydrodynamic coupled solution zone.

[0038] An inlet, outlet, sidewalls, bottom, and top boundary are defined in the extreme wave pool. The inlet is set as the wave-generating plate motion boundary or velocity input boundary, the outlet is set as a wave-damping region, the bottom is set as a wall boundary, and the top is set as a pressure outlet boundary. The sidewalls are set as symmetrical boundaries, slip boundaries, or wall boundaries according to the calculation requirements. To ensure the accuracy of extreme anomalous wave propagation, evolution, and focusing, specific mesh refinement settings are applied near the still water surface, around the hull, and in the coupled solution region.

[0039] Step S2: Create the hull model, liquid tank model, and sloshing plate, and divide the fluid region: like Figure 5 As shown in the figure, this diagram illustrates the cross-sectional dimensions of the hull and the liquid tanks. The diagram displays the cross-sectional outlines of the hull and the liquid tanks located inside the hull. The fixed connection between the liquid tanks and the hull is indicated. The position and height of the sway control plates and the height of the free surface of the liquid are also marked inside the liquid tanks. This step involves constructing a three-dimensional geometric hull model based on the hull lines and main dimensions, and establishing a liquid tank model based on the tank dimensions, tank type, and loading status. The liquid tank model is positioned at a designated location inside the hull and maintains a fixed connection with the hull, achieving the goal of synchronous movement of the liquid tanks with the hull during the calculation process.

[0040] A rigid anti-sway plate model is established inside the liquid tank. This anti-sway plate model can be a vertical anti-sway plate, a horizontal anti-sway plate, a partial anti-sway plate, a perforated anti-sway plate, or a combination thereof. The height, position, opening type, number, and arrangement of the anti-sway plate are adjustable parameters used to construct different anti-sway schemes. The anti-sway plate and the liquid tank wall together constitute the motion boundary of the internal fluid region and are updated synchronously in real time with the ship's movement.

[0041] Set the hull mass, center of gravity position, moment of inertia, initial attitude, initial velocity, and initial angular velocity. Set the filling rate, density, viscosity, and initial free surface height of the liquid inside the tank. Based on calculation requirements, establish comparative working conditions for empty tanks, solid loading, and liquid loading. Solid loading characterizes the state where the liquid mass inside the tank moves with the hull as a whole but does not experience free surface sloshing, while liquid loading characterizes the state where the liquid inside the tank experiences free surface sloshing under the excitation of hull motion. Both conditions have the same mass, center of gravity, and initial moment of inertia.

[0042] like Figure 6 As shown in the figure, the key area mesh settings and partitioning methods are illustrated. The figure shows the spatial distribution relationship of three regions: the external fluid region, the overlapping mesh region, and the internal fluid region. The external fluid region covers the entire computational domain, the overlapping mesh region surrounds the hull and liquid tanks with a closed outline, and the internal fluid region corresponds to the internal space of the liquid tanks. The figure uses different legends to label the distribution locations of the wave-generating mesh, the motion background mesh, the free surface mesh, and the free surface adaptive mesh.

[0043] This step also establishes an external fluid region, an overlapping mesh region, and an internal fluid region. The external fluid region corresponds to the wave propagation region in the extreme wave pool and the external flow field region of the hull, and is used to calculate the propagation, evolution, focusing, diffraction, radiation, and hydrodynamic effects of extreme anomalous waves on the hull. The overlapping mesh region surrounds the hull, liquid tanks, and sway dampers, and moves synchronously with the hull's motion, used to handle the real-time synchronous updating of the mesh during the hull's large-scale motion response. The internal fluid region corresponds to the region containing air and liquid inside the liquid tanks, and is used to calculate and analyze problems such as free surface sloshing, surface rise, liquid impact, bulkhead pressure, and vortex discharge near the sway dampers.

[0044] Both the external and internal fluid regions require the establishment of a two-phase model consisting of a gas and a liquid continuum, and the fluid volume method (VOF) is used to capture the free surface. This method of region division allows for the simultaneous capture of the external wave free surface and the internal sloshing wave free surface within the same viscous flow theory calculation framework. This enables the synchronous action of the ship's motion boundary on the two types of free surfaces, achieving the goal of synchronous full coupling between external anomalous waves and internal sloshing waves.

[0045] Generate the mesh, set the boundary conditions and solution parameters: Computational meshes are generated for the wave field generation and evolution region, the region around the hull, the pre-defined overlapping region, the internal region of the liquid tank, and the region near the oscillation plate. The background mesh covers the entire numerical wave pool, the overlapping mesh surrounds the hull and liquid tank, and the internal mesh is used to analyze the sloshing flow inside the liquid tank. Flow field information is exchanged between the background mesh and the overlapping mesh through interpolation to ensure the continuity of the computational domain during hull motion.

[0046] Local mesh refinement is applied near the free surface in the wave-making area, near the hull surface, near the free surface inside the liquid tanks, near the bulkheads, and near the wave-damping plates. For example... Figure 6 As shown in the figure, the distribution locations of various types of densified grids are marked with different legends. Wave-generating densified grids are distributed near the free surface in the wave-generating area, free-surface densified grids are distributed near the free surface along the wave propagation path, and motion background densified grids are distributed around the hull. Densification in the wave-generating area is used to improve the accuracy of extreme wave evolution calculations, densification around the hull is used to improve the accuracy of hydrodynamic load and motion response calculations, and densification inside the tanks and near the sway-damping plates is used to improve the ability to capture the sloshing wave surface rise, local impact, and vortex energy dissipation processes.

[0047] The time step, number of internal iterations, pressure and velocity coupling algorithm, discretization scheme, turbulence control model, and convergence criteria are set. The time step size is comprehensively analyzed based on the extreme anomaly wave period, the ship's natural motion period, and the characteristic period of the liquid tank sloshing wave, so that the anomaly wave propagation, ship motion, and sloshing wave action process can all be stably analyzed in the time domain. At the same time, combined with the discrete grid size, a high-precision description of the motion of water particles in extreme waves at key locations is achieved.

[0048] Step S3: Establish the reference coordinate system and the ship's coordinate system, and set the motion freedom control conditions: A reference coordinate system fixed to the extreme wave pool is established to describe the propagation of anomalous waves, the pool boundary conditions, and the motion of the ship in global space. A ship-based coordinate system fixed to the ship is established to describe the motion of the tanks, oscillation plates, and internal fluids relative to the ship. The reference and ship-based coordinate systems are converted using ship displacement, rotation angle, linear velocity, and angular velocity to achieve effective correlation of the three physical field information.

[0049] Dynamic fluid-structure interaction (DFBI) motion control conditions are set on the hull model. The hull's degrees of freedom include pitch, sway, heave, roll, pitch, and bow. Based on the response mechanism analysis requirements, various activation states of the degrees of freedom can be configured to form typical working conditions such as single-degree-of-freedom, two-degree-of-freedom, and three-degree-of-freedom. Through the constraint and combination of degree-of-freedom settings, effective separation and coupling calculation of the roll, heave, and sway motion degrees of freedom under extreme anomalous wave excitation can be achieved. This reveals the coupling mechanism of each degree of freedom to the tank sloshing and impact and the amplification of the hull's motion response under extreme anomalous wave excitation conditions.

[0050] Step S4: Generate an abnormal wave with a specified waveform through the initial parameter self-adjustment module. This step generates the initial conditions for the target anomalous wave based on the specified wave height, period, generation location, and focusing time. The target anomalous wave can be generated using a linear physical mechanism or a nonlinear modulation mechanism.

[0051] When generating linear anomalous waves using the new wave theory, the target wave spectrum is discretized into multiple frequency component waves. The frequency, wave number, amplitude, and initial phase of each component wave are determined. Then, through phase adjustment techniques combined with an improved energy-leakage-free Fourier algorithm, the component waves are superimposed at specified locations and times to form an extreme anomalous wave. The generated wavefront elevation is expressed as: ; In the formula, , , and Let be the amplitude, wave number, angular frequency, and corrected phase of the (i)th component wave, respectively. and These represent the target focusing position and focusing time, respectively. By adjusting the amplitude and phase of each component wave, an anomalous wave with a specified height and period is formed at the target position.

[0052] When generating linear anomalous waves using the transient wave packet method, an initial wave train is constructed based on a frequency-sweeping signal whose frequency varies with time, allowing different frequency components to gradually converge during propagation. The wavefront elevation of the transient wave packet is expressed as: ; ; In the formula, This is the distance from the wave-generating position to the theoretical focusing position. The amplitude corresponding to the local frequency. This is a smoothing window function used to reduce abrupt changes in the start and end positions of the wave train. For reference time.

[0053] When anomalous waves are generated using a nonlinear modulation mechanism, the wave energy is concentrated at the target location by controlling the evolution of the carrier wave and wave packet envelope under nonlinear and dispersive effects. The nonlinear wave packet satisfies the following equation, and the free liquid surface is reconstructed from the complex envelope: ; ; In the formula, For nonlinear wave packet complex envelope, For group velocity, and These are the dispersion coefficient and the nonlinear coefficient, respectively. and These are the carrier wave number and angular frequency, respectively. By adjusting the carrier amplitude, initial envelope shape, and propagation distance, the nonlinear wave packet is made to form an anomalous wave with a specified waveform at the target location and target time.

[0054] like Figure 7 As shown in the figure, the three curves illustrate the change in wavefront elevation over time generated by the new wave theory method, the transient wave packet method, and the nonlinear modulation method. The horizontal axis represents time, and the vertical axis represents wavefront elevation. All three curves exhibit a distinct wavefront peak at the target focusing point, indicating that all three methods generate anomalous waves with concentrated energy at the specified location and time, and the peak height and occurrence time are essentially consistent. The new wave theory method, the transient wave packet method, and the nonlinear modulation method generate anomalous waves with similar target wave heights and characteristic periods, respectively, and all three waveforms exhibit significant energy concentration at the target focusing location.

[0055] Specifically, for anomalous waves generated by new wave theory, transient wave packet method, and nonlinear modulation method, the initial parameters are adjusted according to the same target wave height, period, focusing position, and focusing time. The generated initial wave train is input into the extreme wave pool, and the actual evolution waveform is recorded at preset monitoring points and compared with the target anomalous wave waveform. If the monitored waveform meets the target requirements, the corresponding wave-generating input signal is stored; if it does not meet the requirements, the amplitude and phase of the initial wave train are corrected, and the wave evolution calculation is repeated until the preset accuracy requirement is met.

[0056] Step S5, bidirectional fully coupled solution: like Figure 2As shown in the figure, the arrow pointing from the anomalous wave generation module to the external flow field solution module is labeled "Incident Wave Condition"; the arrow pointing from the external flow field solution module to the hull force and motion solution module is labeled "External Wave Pressure"; the arrow pointing from the hull force and motion solution module to the tank sloshing solution module is labeled "Busway Motion Boundary"; the arrow returning from the tank sloshing solution module to the hull force and motion solution module via the sloshing load feedback module is labeled "Sloshing Pressure, Inertial Force, and Torque"; and the arrow pointing from the hull force and motion solution module to the anomalous wave generation module is labeled "Hull Motion Disturbance". This step establishes a coupled solution between anomalous wave evolution excitation, hull motion response analysis, and tank sloshing wave feedback. Figure 2 The information transmission relationship between the modules is shown in detail: the abnormal wave generation module provides the incident wave conditions to the external flow field solution module, the external flow field solution module calculates the external wave pressure of the hull and transmits it to the hull force and motion solution module; the hull motion results are further used as the liquid tank wall motion boundary input to the liquid tank sloshing solution module; the pressure, inertial force and torque generated by the liquid tank sloshing are returned to the hull motion equation through the sloshing load feedback module to update the hull force; after the hull generates a new motion response, it will interfere with the abnormal wave focusing evolution process.

[0057] Within each time step, an anomalous wave is first generated and propagated in the external fluid region, driven by the wave-generating boundary. After the anomalous wave reaches the vicinity of the hull, the transient pressure and viscous shear stress on the outer surface of the hull are solved in the external fluid region, and the external wave excitation force and moment are obtained through surface integration. After the hull moves, the tank walls and sloshing plates move synchronously with the hull. The internal fluid region calculates the pressure and viscous effects generated by tank sloshing and obtains the internal sloshing force and moment. The internal and external fluid loads are expressed as follows: ; ; ; ; In the formula, For the wet surface of the ship's exterior hull, and These are the inner wall of the liquid tank and the surface of the baffle plate, respectively. and These are the stress tensors for the external fluid and the internal fluid, respectively. This indicates the location of the ship's center of gravity.

[0058] The DFBI motion solver module simultaneously inputs external abnormal wave excitation loads and internal sway feedback loads into the ship's motion control equations. The ship's multi-degree-of-freedom motion equations can be uniformly expressed as: ; In the formula, Let be the rigid body mass matrix of the ship's hull and its cargo. Let be the generalized motion vector composed of the ship's displacement and rotation angle. Representing inertial, damped, and viscous hydrodynamic terms, This represents gravity, buoyancy, and still water restoring load. For the generalized excitation load generated by external anomalous waves, The generalized feedback load generated by the sloshing of the liquid tank. The constraint load represents the degree of freedom. This equation shows that the hull motion is jointly controlled by external abnormal wave loads and internal sway loads, rather than being driven solely by external waves.

[0059] After the ship's motion state is updated, the overlapping grid region, the ship's coordinate system, the tank walls, and the sway-damping plate boundaries are updated synchronously. Through the above cycle, a closed-loop two-way coupling is achieved: "external abnormal wave load drives ship motion—ship motion excites tank sloshing—tank sloshing load feeds back to ship motion—updated ship motion again changes the internal and external flow fields."

[0060] like Figures 8 to 12 As shown, the key response time history curves obtained by the fully coupled module are displayed. Among them, Figure 8 The horizontal axis represents time, and the vertical axis represents the roll angle. The curve increases sharply from a small fluctuation to a peak near the moment when the abnormal wave energy accumulates, and then gradually decays. Figure 9 The horizontal axis represents time, and the vertical axis represents heave displacement. The curve exhibits obvious up-and-down oscillation and amplification characteristics near the moment when the abnormal wave energy accumulates. Figure 10 The horizontal axis represents time, and the vertical axis represents sway displacement. The curve shows significant shifts and oscillations near the moment when the abnormal wave energy accumulates. Figure 11 The horizontal axis represents time, and the vertical axis represents the height of liquid level rise. The curve increases sharply and reaches a peak near the moment when the abnormal wave energy accumulates. Figure 12 The horizontal axis represents time, and the vertical axis represents the swaying torque. The curve shows a significant peak fluctuation near the moment when the abnormal wave energy accumulates. Near the moment when the abnormal wave energy accumulates, the amplitudes of the responses of each physical quantity show significant abrupt changes: the ship's roll, heave, and sway motions increase instantaneously, the liquid level in the tank increases synchronously along the walls, and a significant overshoot phenomenon occurs, leading to a peak fluctuation in the swaying torque. This set of curves is used to characterize the synchronous coupling relationship between external abnormal wave input, the ship's three-degree-of-freedom motion, the tank swaying response, and the swaying load feedback.

[0061] Step S6: Extract key variable indicators of the fully coupled model and conduct intelligent optimization analysis of the oscillator based on the KAN surrogate model. The fully coupled solution module based on viscous flow theory is used to obtain the synchronous response results under the interaction of external anomalous waves, hull motion, and tank sloshing waves. Output results include external wavefront elevation, anomalous wave waveform parameters, hull roll angle time history, heave displacement time history, sway displacement time history, tank wall rise time history, bulkhead pressure, sway plate pressure, roll moment, sloshing moment, external flow field velocity vector, internal flow field velocity vector, vorticity, and free surface volume fraction.

[0062] Based on this, the KAN surrogate model is used to establish a nonlinear mapping relationship between the focusing wave parameters, liquid tank parameters, oscillation plate parameters, and the fully coupled response index. The basic mapping relationship of the KAN surrogate model is expressed as follows: ; In the formula, This is the output of the i-th node in the l-th layer. The number of nodes in the l-th layer. A learnable univariate function set on the connecting edges of adjacent nodes.

[0063] In this embodiment, the KAN proxy model includes an input layer, two intermediate layers, and an output layer, and the network node structure is set as follows: ; The input layer has three nodes, each for inputting the period of anomaly spectral peaks. Liquid tank fill rate Relative height of the sway control plate The first intermediate layer has 16 nodes, the second intermediate layer has 12 nodes, and the output layer has 4 nodes, which output the roll amplitude. sag amplitude sway amplitude and liquid level rise .

[0064] The learnable functions on each connecting edge are parameterized using third-order B-spline functions, with the spline order set to 3 and the initial number of grid intervals set to 8. During training, the spline grid is adaptively updated according to the input sample distribution, and the grid position is updated once every 50 training epochs to improve the fitting accuracy within different parameter intervals.

[0065] Both input parameters and output responses were normalized to the minimum-maximum method before training. The KAN model is initially trained using the Adam algorithm with an initial learning rate of 0.001 and a maximum training epoch of 1000 epochs. Subsequently, the L-BFGS algorithm is used for fine-tuning training for no more than 300 epochs. To suppress overfitting, a sparse regularization term and a smoothing regularization term are added to the loss function, with the sparse regularization coefficient set to... The smoothing regularization coefficient is set to Training is stopped when the validation set loss does not decrease for 100 consecutive rounds.

[0066] The fully coupled numerical computation samples were divided into training, validation, and test sets in a 70%, 15%, and 15% ratio, respectively. When the actual sample size is small, a five-fold cross-validation method can be used to evaluate the generalization ability of the KAN model. The mean absolute percentage error of the model test set is no greater than 5%, the maximum relative error is no greater than 10%, and the coefficient of determination is [missing information]. When the value is not less than 0.95, the surrogate model is considered to meet the requirements for predicting the parameters of the oscillator.

[0067] After training, given the abnormal spectral peak period and tank filling rate, the relative height of the damping plate is continuously discretized within a preset range. For example, [the following is a possible interpretation:] Candidate parameters are generated at intervals of 0.01 and input into the KAN surrogate model in sequence to quickly predict the hull motion and tank sloshing response corresponding to each candidate height.

[0068] A comprehensive evaluation function is constructed using the normalized values ​​of the roll, heave, sway, and liquid level rise amplitudes relative to the uncontrolled sway plate condition: To comprehensively evaluate different sway control schemes, a comprehensive evaluation function is constructed using the normalized values ​​of the roll, heave, sway, and liquid level rise amplitudes relative to the condition without sway control: ; in, and These represent the hull roll amplitude under the conditions of controlled sway plates and the baseline condition of uncontrolled sway plates, respectively. and These represent the hull heave amplitudes under controlled and uncontrolled baseline conditions, respectively. and These represent the hull roll amplitudes under controlled and uncontrolled baseline conditions, respectively. and These represent the liquid level rise amplitude on the tank wall under the conditions of controlled sloshing plates and the baseline condition of uncontrolled sloshing plates, respectively; comprehensive evaluation function. The smaller the value, the better the overall suppression effect of the sloshing plate scheme on the coupled response of hull motion and liquid tank sloshing.

[0069] The KAN surrogate model continuously predicts within a preset height range to determine the candidate damping plate height corresponding to the minimum value of the comprehensive evaluation function. This candidate height and its adjacent heights are then re-input into the fully coupled viscous flow model described in steps S2 to S7 for high-precision verification. When the relative error between the KAN prediction value and the fully coupled calculation value meets the preset requirements, the corresponding damping plate height, position, and arrangement are output; when the error does not meet the requirements, new fully coupled calculation results are added to the sample database, and the KAN surrogate model is retrained.

[0070] like Figure 3 The diagram illustrates the intelligent optimization process for sway dampers based on the KAN surrogate model. First, a fully coupled viscous flow model is used to generate samples of hull motion and tank sloshing responses under different abnormal spectral peak periods, tank filling rates, and sway damper parameters. Then, the KAN surrogate model is used to establish a nonlinear mapping relationship between input parameters and response indicators, and to predict the continuous parameter space of the sway damper height. Candidate optimization heights are determined based on a comprehensive evaluation function, and the candidate heights and adjacent heights are then re-input into the fully coupled model for verification. If the prediction error and response control effect meet the requirements, the final optimized sway damper arrangement scheme is output; otherwise, the verification results are added to the sample database, and the surrogate model is retrained.

[0071] like Figure 13 As shown in the figure, the velocity vectors of the external and internal flow fields after adopting the intelligent optimized sway damper scheme are illustrated. The upper part of the figure shows the velocity vector distribution of the external flow field of the hull, and the lower part shows the velocity vector distribution of the internal flow field of the liquid tank. In the external flow field, a dense region of velocity vectors around the hull can be seen, and in the internal flow field, obvious local vortex structures can be seen near the sway damper. The velocity vectors in the vortex region exhibit a rotating distribution pattern, indicating that the optimized sway damper has a hindering and energy-dissipating effect on the large-scale lateral flow of the liquid inside the tank.

[0072] Example 2: This embodiment provides a fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory. The system consists of six cooperating modules in its functional architecture.

[0073] The extreme wave pool module is configured to generate and propagate anomalous waves. The extreme wave pool, along the wave propagation direction, sequentially includes a wave-generating zone, a coupled solution zone, and a wave-dissipating zone. This module incorporates a wave-dissipating damping control algorithm, which automatically adjusts the damping coefficient at the end of the wave-dissipating zone to absorb wave energy and suppress reflected wave interference, ensuring the purity of the wave environment within the coupled solution zone. Simultaneously, this module supports dynamic switching of the inlet boundary type, allowing configuration between the wave-generating plate motion boundary and the velocity input boundary as needed.

[0074] The hull-tank-sway plate model module, configured to be located within the coupled solution area, includes a hull model, tank models located inside the hull, and sway plate models located inside the tanks. This module provides a parametric modeling interface, supporting rapid modification and updating of parameters such as the hull's main dimensions, tank type and size, and sway plate height and position. The module also maintains model data configurations for three comparative operating conditions: empty tank, solid loading, and liquid loading. Solid loading and liquid loading maintain consistency in mass, center of mass, and initial moment of inertia, facilitating comparative analysis of the impact of cargo condition on the coupled response.

[0075] The fluid domain partitioning module is configured to create an external fluid domain, an overlapping mesh domain, and an internal fluid domain. Both the external and internal fluid domains employ the VOF (Volume of Fluid) method to capture the free surface. This module independently configures VOF free surface capture parameters for the external and internal fluid domains, accommodating differences in wave height, period, and surface morphology between external anomalous waves and internal sloshing waves. The module also manages the interpolation weights between the background mesh and the overlapping mesh, dynamically adjusting the interpolation algorithm to ensure computational domain continuity during significant ship movement.

[0076] The anomalous wave generation module is configured to generate anomalous waves with the same target parameters using both linear physical mechanisms and nonlinear modulation mechanisms. The generated initial wave trains are then input into an extreme wave pool for evolution until the target waveform requirements are met. This module incorporates two types of anomalous wave generation algorithms: linear physical mechanisms and nonlinear modulation mechanisms. The linear physical mechanisms include new wave theory algorithms and transient wave packet algorithms, while the nonlinear modulation mechanisms solve wave packet evolution based on the nonlinear Schrödinger equation or its variants. The module includes an initial parameter self-adjusting subunit that automatically compares the waveforms at monitoring points with the target waveform, iteratively correcting the amplitude and phase of the initial wave train until the wave generation accuracy meets a preset threshold. The module also establishes an extreme wave database to store and reuse calibrated wave generation signals.

[0077] The bidirectional fully coupled solution module is configured to generate and propagate anomalous waves in the external fluid region driven by the wave-generating boundary at each time step. It updates the hull motion state in real time based on the external wave loads and the internal sloshing feedback loads of the liquid tanks. After the hull motion state is updated, the overlapping mesh region, liquid tank walls, and sloshing plate boundaries are updated synchronously, achieving a bidirectional fully coupled solution between anomalous wave evolution excitation, hull motion response, and liquid tank sloshing wave feedback. This module incorporates a DFBI solver, simultaneously inputting the external anomalous wave excitation loads and the internal sloshing feedback loads into the hull motion control equations, achieving load superposition in both directions. The module records the number of coupled iterations and convergence residuals at each time step, supporting real-time monitoring and diagnosis of convergence.

[0078] The KAN surrogate model optimization module is configured to construct a KAN surrogate model. It establishes a nonlinear mapping relationship using anomalous wave parameters, liquid tank parameters, and oscillating plate parameters as inputs and a fully coupled response index as output. The KAN surrogate model is used to continuously predict and optimize the oscillating plate parameters. Candidate optimization schemes are then re-input into the bidirectional fully coupled solution module for verification, outputting a satisfactory optimized oscillating plate arrangement. This module includes a sample management subunit for cleaning, normalizing, and maintaining the sample library of fully coupled calculation results; a network training subunit for executing a hybrid Adam and L-BFGS training strategy and monitoring the loss curves of the training and validation sets to prevent overfitting; a prediction and optimization subunit for rapid prediction in the continuous space of oscillating plate parameters and automatically determining candidate optimization parameters using a comprehensive evaluation function; and a verification management subunit for submitting candidate optimization schemes to the bidirectional fully coupled solution module for verification calculations and comparing the relative error between the predicted and verified values ​​to determine whether to output the final scheme or trigger sample supplementation and model retraining.

[0079] The modules described above are decoupled and collaborate through standardized data interfaces, supporting independent upgrades and replacements at the module level. The system can automatically execute a complete closed-loop process, from anomalous wave generation and fully coupled simulation to oscillator parameter optimization, without human intervention.

[0080] Example 3: This embodiment provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in Embodiment 1.

[0081] The computer-readable storage medium can be implemented in a variety of physical forms.

[0082] In one implementation, the medium is a solid-state drive or a hard disk drive, deployed in a high-performance computing workstation or server, used to store mesh files, boundary condition configurations, and solution parameter settings during the fully coupled simulation process.

[0083] As another implementation, the medium is a portable storage device used to transfer calibrated anomalous wave generation signal databases and trained KAN surrogate model parameter files between different computing platforms.

[0084] As another implementation method, the medium is a cloud storage node, which enables multiple computing nodes to share access to the sample database and model parameters through a distributed file system, supporting parallel computing of fully coupled simulations and large-scale parameter optimization.

[0085] After being loaded by the processor, the computer program executes the steps described in Example 1 in sequence: first, it initializes the extreme wave pool and the hull-liquid tank-steamer model; then, it divides the fluid region and sets boundary conditions; next, it calls the anomalous wave generation module to perform initial wave train calibration; then, it enters a fully coupled time-stepping loop, solving the external flow field, hull motion, and internal flow field sequentially in each time step and updating the coupled boundaries; finally, after completing the fully coupled simulation, it extracts response indices and drives the KAN proxy model to optimize the steamer parameters. The computer program also includes an anomaly handling subroutine, which automatically adjusts the time step size or refines the local mesh and re-executes the calculation for that time step when the coupled iteration in a certain time step fails to converge.

[0086] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A fully coupled simulation system for anomalous waves and sloshing waves in liquid cargo ships based on viscous flow theory, characterized in that, include: An extreme wave pool module is configured to generate and propagate abnormal waves. The extreme wave pool includes a wave-generating zone, a coupled solution zone, and a wave-dissipating zone in sequence along the wave propagation direction. The hull-liquid tank-squeezing plate model module is located within the coupled solution area and includes a hull model, a liquid tank model located inside the hull, and a squeezing plate model located inside the liquid tank. The fluid region partitioning module is configured to create an external fluid region, an overlapping grid region, and an internal fluid region. Both the external and internal fluid regions use the fluid volume method to capture the free liquid surface. The abnormal wave generation module is configured to generate abnormal waves with the same target parameters using linear physical mechanisms and nonlinear modulation mechanisms respectively, and input the generated initial wave train into the extreme wave pool for evolution until the target waveform requirements are met. The bidirectional fully coupled solution module is configured to generate and propagate abnormal waves in the external fluid region driven by the wave-generating boundary at each time step. The hull motion state is updated in real time according to the external wave load and the internal sloshing feedback load of the liquid tank. After the hull motion state is updated, the overlapping grid region, the liquid tank wall and the sloshing plate boundary are updated synchronously, realizing the bidirectional fully coupled solution between the abnormal wave evolution excitation, the hull motion response and the liquid tank sloshing wave feedback. The KAN surrogate model optimization module is configured to construct a KAN surrogate model. It establishes a nonlinear mapping relationship with abnormal wave parameters, liquid tank parameters, and oscillation plate parameters as inputs and a fully coupled response index as output. The KAN surrogate model is used to continuously predict and optimize the oscillation plate parameters. The candidate optimization schemes are re-input into the bidirectional fully coupled solution module for verification, and the optimized oscillation plate arrangement scheme that meets the requirements is output.

2. The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory according to claim 1, characterized in that, In the anomalous wave generation module, the linear physical mechanism for generating anomalous waves includes at least one of the new wave theory method and the transient wave packet method. The new wave theory method discretizes the target wave spectrum into multiple frequency component waves, determines the frequency, wave number, amplitude, and initial phase of each component wave, and uses phase adjustment technology to superimpose each component wave at a specified position and time to form an anomalous wave. The transient wave packet method constructs an initial wave train based on a frequency sweep signal that varies with time, so that different frequency components gradually converge to form an anomalous wave during propagation. The nonlinear modulation mechanism for generating anomalous waves is based on nonlinear wave theory or nonlinear envelope wave method, and controls the evolution of the carrier and wave packet envelope under nonlinear and dispersion effects to concentrate wave energy at the target position to form an anomalous wave.

3. The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory according to claim 1, characterized in that: In the fluid region partitioning module, the external fluid region is configured to calculate the propagation, evolution, focusing, diffraction, radiation, and hydrodynamic effects of anomalous waves on the hull; the overlapping grid region is configured to surround the hull, liquid tanks, and sway dampers and move synchronously with the hull movement to handle real-time grid updates during hull movement; the internal fluid region is configured to calculate free surface sloshing, surface rise, liquid impact, bulkhead pressure, and vortex discharge near the sway dampers inside the liquid tanks; both the external and internal fluid regions are configured with a gas and liquid dual-continuum two-phase model.

4. The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory according to claim 1, characterized in that: It also includes a coordinate system and motion control module, configured to establish a reference coordinate system fixed in the extreme wave pool and a ship-mounted coordinate system fixed in the hull, and set dynamic fluid-structure interaction motion solution control conditions; the hull motion degrees of freedom include pitching, swaying, heeling, rolling, pitching and bowing, and by setting the activation state of the degrees of freedom, single-degree-of-freedom, double-degree-of-freedom and multi-degree-of-freedom working conditions are formed to realize the separation and coupling calculation of each motion degree of freedom.

5. The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory according to claim 1, characterized in that: In the bidirectional fully coupled solution module, the information transmission relationship between the abnormal wave evolution excitation, the hull motion response, and the feedback effect of the liquid tank sloshing wave is expressed as follows: ； ； in, Let the generalized motion vector of the ship's hull be . and These represent the external flow field and the internal flow field states, respectively. For the generalized excitation load generated by external anomalous waves, The generalized feedback load generated by the sloshing of the liquid tank. This represents the process of solving for the ship's motion. This represents the process of updating the moving boundary and the flow field. This represents the number of coupling iterations within the same physical time step.

6. The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory according to claim 1, characterized in that: In the KAN proxy model optimization module, the basic mapping relationship of the KAN proxy model is expressed as follows: ； in, For network layer number, and The first Layer and first Layer node number, This is the output of the i-th node in the l-th layer. The number of nodes in the l-th layer. To connect the first Layer The node and the first Layer The KAN surrogate model has learnable univariate functions for each node; the input parameters of the KAN surrogate model include at least the period of the abnormal spectral peak, the liquid tank filling rate, and the relative height of the oscillation plate, and the output parameters include at least the roll amplitude, heave amplitude, sway amplitude, and liquid level rise amplitude; the network node structure of the KAN surrogate model is set to 3-16-12-4, that is, the input layer has 3 nodes, the first intermediate layer has 16 nodes, the second intermediate layer has 12 nodes, and the output layer has 4 nodes; the learnable univariate functions on each connection edge are parameterized using third-order B-spline functions.

7. The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory according to claim 1, characterized in that: In the KAN surrogate model optimization module, when continuously predicting and optimizing the damper parameters using the KAN surrogate model, a comprehensive evaluation function is used to evaluate the candidate damper schemes: ； in, and These represent the hull roll amplitude under the conditions of controlled sway plates and the baseline condition of uncontrolled sway plates, respectively. and These represent the hull heave amplitudes under controlled and uncontrolled baseline conditions, respectively. and These represent the hull roll amplitudes under controlled and uncontrolled baseline conditions, respectively. and These represent the liquid level rise amplitude on the tank wall under the conditions of controlled sloshing plates and the baseline condition of uncontrolled sloshing plates, respectively; comprehensive evaluation function. The smaller the value, the better the overall suppression effect of the sway plate scheme on the coupled response of hull motion and liquid tank sloshing.

8. The fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory according to claim 1, characterized in that: In the KAN surrogate model optimization module, when the candidate optimization scheme is re-input into the fully coupled viscous flow model for verification, if the relative error between the predicted value of the KAN surrogate model and the calculated value of the fully coupled model meets the preset requirements, the corresponding damping plate height, position and arrangement will be output. If the error does not meet the requirements, the new fully coupled calculation results will be added to the sample database and the KAN surrogate model will be retrained.

9. A fully coupled simulation method for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory, applied to the fully coupled simulation system for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory as described in any one of claims 1-8, characterized in that, include: S1: Establish an extreme wave pool model, which includes a wave-generating zone, a coupled solution zone, and a wave-dissipating zone along the wave propagation direction. S2: Establish a hull model, a liquid tank model, and a sway plate model within the coupled solution region, and establish an external fluid region, an overlapping mesh region, and an internal fluid region. The external fluid region and the internal fluid region both use the fluid volume method to capture the free liquid surface. S3: Establish a reference coordinate system fixed in the extreme wave pool and a ship-mounted coordinate system fixed in the hull, and set dynamic fluid-structure interaction motion solution control conditions. The hull motion degrees of freedom include pitching, swaying, heaving, rolling, pitching and bowing. By setting the activation state of the degrees of freedom, single-degree-of-freedom, double-degree-of-freedom and multi-degree-of-freedom working conditions are formed to realize the separation and coupling calculation of each motion degree of freedom. S4: Abnormal waves with the same target parameters are generated using linear physical mechanisms and nonlinear modulation mechanisms respectively, and the generated initial wave train is input into the extreme wave pool for evolution until the target waveform requirements are met; S5: In each time step, the external fluid region is driven by the wave-generating boundary to generate and propagate abnormal waves. The hull motion state is updated in real time according to the external wave load and the internal sloshing feedback load of the liquid tank. After the hull motion state is updated, the overlapping grid region, the liquid tank wall and the sloshing plate boundary are updated synchronously to realize the bidirectional full coupling solution between the abnormal wave evolution excitation, the hull motion response and the liquid tank sloshing wave feedback. S6: Construct a KAN surrogate model, using anomalous wave parameters, liquid tank parameters, and oscillation plate parameters as inputs and the fully coupled response index as output to establish a nonlinear mapping relationship. Use the KAN surrogate model to continuously predict and optimize the oscillation plate parameters. Re-input the candidate optimization scheme into the fully coupled viscous flow model for verification and output an optimized oscillation plate arrangement scheme that meets the requirements.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the fully coupled simulation method for anomalous waves and sloshing waves of liquid cargo ships based on viscous flow theory as described in claim 9.

Images