Thick plate welding residual stress numerical prediction method based on heat-fluid-solid multi-field coupling

Abstract

The invention discloses a thick plate welding residual stress numerical prediction method based on heat-fluid-solid multi-field coupling, which combines computational fluid dynamics with a finite element method, and comprehensively considers the flow characteristic and thermodynamic effect of liquid metal. In the CFD, the influence of metal melting, molten pool flowing, surface tension, buoyancy, heat transfer and fluid flowing on the geometric shape and temperature distribution of the weld joint is simulated, in the FEM module, the geometric shape and temperature difference of the weld joint are calculated based on the CFD, residual stress and deformation are further simulated and analyzed in combination with a life-death unit technology, and through the FEM module, the residual stress and deformation of the weld joint are obtained. The forming mechanism of the convex welding seam, the evolution rule of the thick plate welding residual stress and the influence of the geometrical shape on residual stress concentration are systematically researched.

Description

Technical Field

[0001] The invention belongs to the field of simulation calculation of residual stress in thick plate welding, and in particular relates to a numerical prediction method for residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling. Background Art

[0002] In recent years, low-alloy high-strength steel has been widely used in many fields such as marine engineering due to its high strength, good toughness, corrosion resistance and good weldability. Welding is a connection process that uses local high-temperature heating to fill the weld with molten filler metal and achieve a firm connection with the parent material. However, during the welding process, due to the influence of thermal cycles, uneven temperature distribution and thermal gradients, the microstructure of the welded parts will change, resulting in residual stress and deformation, which will significantly affect its mechanical properties. In addition, with the large-scale structure of offshore platforms, the thickness of high-strength steel plates has increased significantly, and multi-layer and multi-pass butt welding is usually used. Due to repeated rapid heating and cooling, a more complex welding residual stress distribution is formed inside the butt joint, making it a weak link in the mechanical properties of the welded structure.

[0003] As a basic analysis tool in the field of welding, the traditional thermo-mechanical model has become the mainstream method for simulating the evolution of temperature field, residual stress distribution and component deformation prediction during welding with its relatively simple theoretical framework, providing an important theoretical basis for the optimization of welding parameters in engineering practice. However, although traditional thermodynamic analysis can predict welding residual stress to a certain extent, its inherent limitation is that it cannot consider the important influence of molten pool flow behavior on temperature field distribution and weld morphology. In addition, the modeling process usually relies on a pre-assumed weld bead profile or assumes an idealized flat surface, which makes it difficult to truly reflect the complex physical phenomena in the welding process. Summary of the invention

[0004] The purpose of the present invention is to overcome the above shortcomings and provide a method for numerically predicting residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling.

[0005] The purpose of the present invention is achieved through the following technical solutions: A numerical prediction method for thick plate welding residual stress based on thermal-fluid-solid multi-field coupling, comprising:

[0006] S1. Heat source model construction and parameter definition;

[0007] S2, CFD testing and thermal flow field simulation verification;

[0008] S3, FEM module residual stress prediction;

[0009] S4, multi-physics field coordination and parameter adjustment;

[0010] S5, experimental verification and model modification;

[0011] S6. Output and application of process plan.

[0012] A further improvement of the present invention is that the specific steps include:

[0013] S1. Heat source model construction and parameter definition: According to the welding process requirements, a heat source model is established, and the three-dimensional distribution of heat input during welding is simulated through the CFD module. In combination with the material thermophysical properties, the initial heat source parameters are raised;

[0014] S2. CFD test and thermal flow field simulation verification: Perform CFD analysis on the heat source model in step S1 to simulate the flow behavior, heat flow distribution and temperature field evolution of the welding molten pool; verify the accuracy of the CFD simulation results through experimental measurement and optimize the heat source model parameters;

[0015] S3, FEM module residual stress prediction: Based on the optimized heat source model, the FEM module is used to simulate the three-dimensional temperature field during welding, calculate the residual stress distribution of the welded joint, and output the stress concentration factor and deformation prediction results of the key area;

[0016] S4. Multi-physics field collaborative optimization and parameter adjustment: Based on the CFD thermal flow field distribution and FEM residual stress prediction results, the welding process parameters are adjusted to optimize the heat source model to ensure that the thermal-mechanical coupling performance meets the design requirements;

[0017] S5, experimental verification and model correction: prepare actual welding samples, measure the residual stress distribution by blind hole method or X-ray diffraction method, compare the FEM prediction results of step S3, and iteratively correct the simulation model according to the experimental data;

[0018] S6. Process plan output and application: Integrate the optimized heat source model parameters and FEM residual stress distribution map to form a welding process guidance plan, clarify key indicators such as heat input control range and residual stress safety threshold, and directly apply it to the thick plate welding process design and quality assessment of offshore structures.

[0019] A further improvement of the present invention is that in CFD, the flow of the welding pool follows the three conservation laws of fluid motion: the law of conservation of mass, the law of conservation of momentum and the law of conservation of energy, and their expressions are as follows:

[0020] (1) ;

[0021] (2) ;

[0022] (3) ;

[0023] In the formula, , , , , , , , , , They are material density, time, velocity, mass source term, static pressure, viscosity, momentum source term, material enthalpy, thermal conductivity, and energy source term. The material enthalpy expression is as follows:

[0024] (4) ;

[0025] (5) ;

[0026] In the formula, , , , , , , They are reference material enthalpy, reference temperature, specific heat capacity, latent heat, liquid volume fraction, solidus temperature, and liquidus temperature;

[0027] In the CFD of step S2, the mushy zone is regarded as a porous medium to avoid the complex calculation of directly tracking the interface between phases. The momentum source term expression is as follows:

[0028] (6) ;

[0029] In the formula, is the paste parameter, is the drag speed, is 0.0001.

[0030] A further improvement of the present invention is that in the CFD of step S2, the surface tension is negatively correlated with the temperature, and its expression is as follows:

[0031] (7) ;

[0032] In the formula, is the surface tension gradient, is the melting point, is the surface tension at the melting point.

[0033] A further improvement of the present invention is that in the CFD of step S2, the buoyancy expression generated by the welding pool is as follows:

[0034] (8) ;

[0035] In formula (8), is the coefficient of thermal expansion;

[0036] The VOF value of the free surface of the molten pool is tracked at each time step by the fluid volume method, and its expression is as follows:

[0037] (9) ;

[0038] In formula (9), represents the fluid volume fraction;

[0039] In the welding process, in addition to heat input, the effects of convection heat transfer and radiation heat transfer on heat flux must also be considered comprehensively. The expression is as follows:

[0040] (10) ;

[0041] In formula (10), is the heat flux density, is the convection heat coefficient, is the material emissivity, is the Stefan-Boltzmann constant.

[0042] A further improvement of the present invention is that in CFD, the heat source model adopts a semi-elliptical volume heat source form, which can effectively approximate the shape and size of the welding pool and accurately simulate the distribution of the mobile heat source during the welding process. The heat source expressions of the front and rear are as follows:

[0043] (11) ;

[0044] (12) ;

[0045] In the formula, , , is the local coordinate system of the heat source model, , , , is the shape parameter of the heat source model, , , For welding efficiency, voltage and current, , is the front and backup heat fraction of the mobile heat source.

[0046] A further improvement of the present invention is that in the FEM module of step S3, the material behavior in the plastic zone satisfies the plastic flow and strain hardening laws, and the total strain is divided into the following three components: (13) ;

[0047] In the formula, is the elastic strain, is the plastic strain, For thermal strain.

[0048] A further improvement of the present invention is that: in CFD, the weld geometry is determined by the VOF method, where the function f represents the volume fraction of the metal term;

[0049] In this module, the following three cases are considered: (1) when f>0.5, it represents the metal item; (2) when f=0.5, it represents the interface between the metal item and the gas item; (3) when f<0.5, it represents the gas item.

[0050] A further improvement of the present invention is that in CFD, the temperature at the center of the Euler control volume in CFD is mapped to the Lagrangian unit node of the FEM module by three-dimensional interpolation, thereby realizing the loading of temperature distribution, and the expression is as follows:

[0051] (14) ;

[0052] In the formula, , , is the local coordinate system of the CFD model, , , are the natural coordinates of the CV centers around the FEM model nodes, is the temperature value of the 8 CV centers adjacent to the unit node, is the interpolation function, is the temperature mapped to the FEM model.

[0053] A further improvement of the present invention is that: by using formula (14), the two adjacent moments in the FEM model are obtained: , Temperature and Afterwards, when Temperature between The approximate temperature is calculated by linear interpolation. The expression is as follows:

[0054] (15) ;

[0055] In order to reasonably and accurately apply the temperature distribution data provided by the CFD model, the time step of the FEM model should not be larger than the time interval of the CFD model, so as to achieve accurate interpolation between adjacent temperature distributions.

[0056] Compared with the prior art, the present invention has the following advantages:

[0057] The present invention combines computational fluid dynamics with the finite element method, and comprehensively considers the flow characteristics of liquid metal and its thermodynamic effects. In CFD, the effects of metal melting, molten pool flow, surface tension, buoyancy, heat transfer, and fluid flow on weld geometry and temperature distribution are simulated. In the FEM module, the weld geometry and temperature difference obtained by CFD calculation are further simulated and analyzed in combination with the birth-death unit technology. Through the FEM module, the formation mechanism of convex welds, the evolution law of residual stress in thick plate welding, and the effect of geometry on residual stress concentration are systematically studied. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the process of the present invention.

[0059] Figure 2 This is a model diagram of double moving heat sources in the present invention.

[0060] Figure 3 Schematic diagram of temperature interpolation between CV center and FEM node in CFD of the present invention.

[0061] Figures 4 to 6 These are the simulation results of the temperature field and flow field of the CFD model on the back and front sides of the weld when the molten pool is in a stable state. DETAILED DESCRIPTION

[0062] In order to deepen the understanding of the present invention, the present invention will be further described in detail below in conjunction with embodiments and drawings. The embodiments are only used to explain the present invention and do not constitute a limitation on the protection scope of the present invention.

[0063] Numerical prediction method of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling, refer to Figure 1 , the specific steps include,

[0064] S1. Heat source model construction and parameter definition: According to the welding process requirements, a heat source model is established, and the three-dimensional distribution of heat input during welding is simulated through the CFD module. In combination with the material thermophysical properties, the initial heat source parameters are raised;

[0065] S2. CFD test and thermal flow field simulation verification: Perform CFD analysis on the heat source model in step S1 to simulate the flow behavior, heat flow distribution and temperature field evolution of the welding molten pool; verify the accuracy of the CFD simulation results through experimental measurement and optimize the heat source model parameters;

[0066] S3, FEM module residual stress prediction: Based on the optimized heat source model, the FEM module is used to simulate the three-dimensional temperature field during welding, calculate the residual stress distribution of the welded joint, and output the stress concentration factor and deformation prediction results of the key area;

[0067] S4. Multi-physics field collaborative optimization and parameter adjustment: Based on the CFD thermal flow field distribution and FEM residual stress prediction results, the welding process parameters are adjusted to optimize the heat source model to ensure that the thermal-mechanical coupling performance meets the design requirements;

[0068] S5, experimental verification and model correction: prepare actual welding samples, measure the residual stress distribution by blind hole method or X-ray diffraction method, compare the FEM prediction results of step S3, and iteratively correct the simulation model according to the experimental data;

[0069] S6. Process plan output and application: Integrate the optimized heat source model parameters and FEM residual stress distribution map to form a welding process guidance plan, clarify key indicators such as heat input control range and residual stress safety threshold, and directly apply it to the thick plate welding process design and quality assessment of offshore structures.

[0070] The present invention combines computational fluid dynamics with the finite element method, and comprehensively considers the flow characteristics of liquid metal and its thermodynamic effects. In CFD, the effects of metal melting, molten pool flow, surface tension, buoyancy, heat transfer, and fluid flow on weld geometry and temperature distribution are simulated. In the FEM module, the weld geometry and temperature difference obtained by CFD calculation are further simulated and analyzed in combination with the birth-death unit technology. Through the FEM module, the formation mechanism of convex welds, the evolution law of residual stress in thick plate welding, and the effect of geometry on residual stress concentration are systematically studied.

[0071] About CFD

[0072] The CFD software fluent is used to simulate the welding process of heat transfer and fluid flow in gas shielded welding and submerged arc welding. In CFD, it is assumed that the fluid in the molten pool is an incompressible Newtonian viscous fluid, and the flow pattern is laminar flow. The flow of the molten pool follows the three conservation laws of fluid motion: the law of conservation of mass, the law of conservation of momentum, and the law of conservation of energy. Their expressions are as follows:

[0073] (1) ;

[0074] (2) ;

[0075] (3) ;

[0076] In the formula, , , , , , , , , , They are material density, time, velocity, mass source term, static pressure, viscosity, momentum source term, material enthalpy, thermal conductivity, and energy source term. The material enthalpy expression is as follows:

[0077] (4) ;

[0078] (5) ;

[0079] In the formula, , , , , , , They are reference material enthalpy, reference temperature, specific heat capacity, latent heat, liquid volume fraction, solidus temperature, and liquidus temperature respectively.

[0080] The present invention uses the Enthalpy-Porosity technology to process the melting and solidification process, and regards the mushy zone as a porous medium, thereby avoiding the complex calculation of directly tracking the interface between phases. The momentum source term expression is as follows:

[0081] (6) ;

[0082] In the formula, is the paste parameter, is the drag speed, is 0.0001.

[0083] Since the maximum temperature of submerged arc welding is not enough to reach the vapor phase temperature of S355 steel, the fluid behavior in the molten pool during submerged arc welding mainly considers the influence of gravity, thermal buoyancy and surface tension.

[0084] Surface tension plays an important role in the flow of molten pool fluid. The negative correlation between surface tension and temperature is shown in the following expression:

[0085] (7) ;

[0086] In the formula, is the surface tension gradient, is the melting point, is the surface tension at the melting point.

[0087] During the welding process, the metal density will change with the temperature. Due to the uneven temperature distribution in the molten pool, the density difference will cause the buoyancy. The buoyancy expression generated by the welding pool is as follows:

[0088] (8) ;

[0089] In formula (8), is the coefficient of thermal expansion;

[0090] The VOF value of the free surface of the molten pool is tracked at each time step by the fluid volume method, and its expression is as follows:

[0091] (9) ;

[0092] In formula (9), represents the fluid volume fraction;

[0093] In the welding process, in addition to heat input, the effects of convection heat transfer and radiation heat transfer on heat flux must also be considered comprehensively. The expression is as follows:

[0094] (10) ;

[0095] In formula (10), is the heat flux density, is the convection heat coefficient, is the material emissivity, is the Stefan-Boltzmann constant.

[0096] In CFD, the double ellipsoid moving heat source model is widely used in the numerical simulation of metal welding processes, such as submerged arc welding and gas shielded welding, etc. Figure 2 The heat source model adopts a semi-elliptical volume heat source form, which can effectively approximate the shape and size of the welding pool and accurately simulate the distribution of mobile heat sources during welding. The front and rear heat source expressions are as follows:

[0097] (11) ;

[0098] (12) ;

[0099] In the formula, , , is the local coordinate system of the heat source model, , , , is the shape parameter of the heat source model, , , For welding efficiency, voltage and current, , is the front and backup heat fraction of the mobile heat source.

[0100] About the FEM Module

[0101] The FEM model uses Mechanical to simulate the subsequent mechanical effects. The FEM model is modeled based on the weld geometry simulated by the CFD model, and the temperature field simulated by the CFD model is used as an external load to calculate the stress field. When calculating the stress field, fixed constraints are applied to the weld plate model to eliminate rigid body motion. In mechanical analysis, temperature-related material properties such as Young's modulus, yield strength, Poisson's ratio and thermal expansion coefficient play a key role. The yield condition of the material follows the Von Mises criterion, and the material behavior in the plastic zone satisfies the laws of plastic flow and strain hardening. The total strain is divided into the following three components:

[0102] (13) ;

[0103] In the formula, is the elastic strain, is the plastic strain, For thermal strain.

[0104] Based on the CFD simulation results, it is necessary to extract the temperature field and the geometry of the weld after welding for the subsequent FEM module thermoelastic-plastic analysis. First, the geometry of the weld is obtained from the CFD data. In CFD, the geometry of the weld is determined by the VOF method, where the function f represents the volume fraction of the metal term; in this module, the following three cases are considered: (1) when f>0.5, it represents the metal term; (2) when f=0.5, it represents the interface between the metal term and the gas term; (3) when f<0.5, it represents the gas term.

[0105] The temperature distribution is loaded by mapping the temperature at the center of the Euler control volume in CFD to the Lagrangian unit node of the FEM module through three-dimensional interpolation, referring to Figure 3 , the expression is as follows:

[0106] (14) ;

[0107] In the formula, , , is the local coordinate system of the CFD model, , , are the natural coordinates of the CV centers around the FEM model nodes, is the temperature value of the 8 CV centers adjacent to the unit node, is the interpolation function, is the temperature mapped to the FEM model.

[0108] The two adjacent moments in the FEM model are obtained by formula (14): , Temperature and Afterwards, when Temperature between The approximate temperature is calculated by linear interpolation. The expression is as follows:

[0109] (15) ;

[0110] In order to reasonably and accurately apply the temperature distribution data provided by the CFD model, the time step of the FEM model should not be larger than the time interval of the CFD model, so as to achieve accurate interpolation between adjacent temperature distributions.

[0111] The model constructed by the present invention includes

[0112] CFD module, which simulates the welding process of heat transfer and fluid flow of gas shielded welding and submerged arc welding through CFD module;

[0113] FEM module, the FEM module models the weld geometry simulated by the CFD module and uses the temperature field simulated by the CFD model as an external load to calculate the stress field;

[0114] Mapping and 3D interpolation, based on the simulation results of the CFD module, extract the temperature field and the geometry of the weld bead after welding for subsequent thermal elastic-plastic analysis of the FEM module to obtain residual stress and strain.

[0115] The present invention constructs a CFD-FEM model to study the welding residual stress in the process of thick plate welding in marine engineering. The model combines CFD and FEM methods, and fully considers the flow characteristics of liquid metal and its thermodynamic effects. In CFD, the effects of metal melting, molten pool flow, surface tension and thermal buoyancy on the geometry and temperature distribution of the weld are simulated. The finite element module is based on the thermoelastic theory and uses the temperature field and geometric field data obtained by CFD simulation to further calculate the welding residual stress.

[0116] Calculation results

[0117] Figures 4 to 6The simulation results of the temperature field and flow field of the CFD model on the back and front of the weld when the molten pool is in a stable state, and the corresponding test result diagram. In the calculation process, the molten pool state is determined by the material temperature performance of S355. When the material temperature exceeds the solidus temperature of S355, the material will be regarded as being in the liquidus state. The red area in the figure is the molten pool. From the top view of the weld surface simulated by CFD, it can be seen that the molten pool is in the shape of a water drop. The tail edge angle of the molten pool on the back is 37.4°, the tail edge angle of the molten pool of the first weld on the front is 38.2°, and the tail edge angle of the molten pool of the second weld on the front is 33.1°. Since the temperature gradient at the center line of the molten pool is the lowest, the solidification rate in this area reaches the maximum value, and the latent heat release rate also reaches the peak value. However, when the maximum solidification rate of the molten pool is lower than the welding speed, the minimum angle between the maximum thermal gradient direction in the tail area and the welding speed fails to reach 0. Affected by this, the heat dissipation efficiency in the tail area is significantly reduced, and the minimum temperature gradient cannot release the latent heat quickly and effectively, resulting in a slowdown in the solidification rate of the liquid metal in the tail area, and ultimately causing the molten pool to present a typical water droplet-like characteristic morphology.

[0118] The convex weld is the result of the combined action of multiple complex factors. From the top view and longitudinal view, it can be observed that the molten metal flows to the rear of the welding pool and gradually accumulates, eventually forming the swelling of the tail. As the welding heat source moves, the size of the tail expansion further increases, and the liquid channel connecting the expansion and the rear of the molten pool becomes slender. When the liquid channel is far away from the arc heat source, the molten metal inside it is prone to rapid solidification due to the lack of sufficient heat support. This premature solidification phenomenon marks the formation of the hump. The solidified hump hinders the continued flow of the molten metal, thereby strengthening the morphology of the hump. In addition, the effect of surface tension is crucial to the formation of the hump. Since the molten metal temperature in the swelling area of ​​the tail is lower than that in the front of the molten pool, and the surface tension temperature coefficient is negative, the low-temperature molten metal has a higher surface tension, which further promotes the growth of the swelling. At the same time, the molten metal temperature drops and the fluidity weakens due to heat loss in the liquid channel, which also increases the tendency of premature solidification of the liquid channel. The study also showed that the bulge phenomenon is more significant when the ratio of the molten pool width to length is small. The causes of convex welds can be summarized into two main factors. First, the high momentum of the backward fluid flow leads to the formation and continued growth of the tail expansion because the amount of filler metal in the molten pool is limited, and the amount of molten metal and heat energy in the liquid channel are also insufficient. Second, the change of the Marangoni effect in the liquid channel intensifies the surface tension gradient in the welding direction, further promoting the contraction of the liquid channel. These two factors together limit the backfilling capacity of the molten metal, making the liquid channel more prone to premature solidification, and ultimately forming a convex weld.

[0119] In multi-pass welding, the regularity of the weld pool flow field is not only affected by the welding heat input, but also by the change in weld bead width. Figure 4 In the back welding (Pass 3), due to the narrow weld bead and concentrated flow field, the molten pool fluid mainly flows along the welding direction, showing high stability. Figure 5 The first pass (Pass 6) is wider, and the increase in its width leads to a more diffuse distribution of the fluid at the rear end of the molten pool, and the fluid tends to expand to both sides, but the overall flow field still maintains a certain regularity. Figure 6 , in the second pass (Pass 7), due to the influence of the sixth weld, the width of the seventh weld is significantly narrowed. At the same time, the locally higher geometric structure of the sixth weld interferes with the molten pool fluid, resulting in a weakening of the regularity of the flow field and a decrease in stability. Part of the fluid is affected by the edge of the sixth weld at the rear end of the molten pool, deviates from the mainstream direction and flows toward the edge. This shows that the change in weld width has a significant effect on the regularity and stability of the molten pool fluid flow. At the same time, the geometric characteristics of adjacent welds, such as the raised area of ​​the sixth weld, also have a certain effect on the flow field distribution. In summary, in multi-pass welding, weld width is one of the key factors affecting the molten pool flow field, and the geometric characteristics of adjacent welds also have an important effect on the flow field.

[0120] The advantages of the CFD-FEM model constructed by the present invention are as follows:

[0121] 1. Accurate analysis of molten pool flow on residual stress: The CFD-FEM model constructed in the present invention not only considers the influence of welding heat source on welding residual stress, but also further considers the influence of flow separation and dynamic characteristics in the molten pool, and reveals the influence of molten pool flow on heat conduction, welding residual stress distribution and deformation.

[0122] 2. Accurate prediction of residual stress by weld bead geometry: The CFD-FEM model constructed in the present invention fully considers the influence of weld bead geometry on welding residual stress during welding, and combined with weld bead geometry factors, it can more accurately predict and analyze the residual stress and strain distribution, especially after thick plate welding.

[0123] 3. Analysis of the correlation between welding defects and process parameters: Welding defects are closely related to process parameters. For example, the settings of voltage, current, welding speed, etc. directly affect the weld formation and quality. By deeply studying the relationship between process parameters and welding defects, it is possible to effectively prevent the occurrence of common defects such as pores, lack of fusion, cracks, and burn-through, and further improve the quality and reliability of the welding process.

[0124] 4. Wide industrial applicability of the model: The CFD-FEM model constructed by the present invention shows significant applicability in a variety of welding materials and processes. It is suitable for processes such as gas shielded welding, submerged arc welding, and laser welding. It can also be applied to welding of the same steel and welding of different materials, and has broad industrial application prospects.

[0125] Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A numerical prediction method for residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling, characterized by: include, S1. Heat source model construction and parameter definition; S2, CFD testing and thermal flow field simulation verification; S3, FEM module residual stress prediction; S4, multi-physics field coordination and parameter adjustment; S5, experimental verification and model modification; S6. Output and application of process plan.

2. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 1 is characterized in that: The specific steps include: S1. Heat source model construction and parameter definition: According to the welding process requirements, a heat source model is established, and the three-dimensional distribution of heat input during welding is simulated through the CFD module. In combination with the material thermophysical properties, the initial heat source parameters are raised; S2. CFD test and thermal flow field simulation verification: Perform CFD analysis on the heat source model in step S1 to simulate the flow behavior, heat flow distribution and temperature field evolution of the welding molten pool; verify the accuracy of the CFD simulation results through experimental measurement and optimize the heat source model parameters; S3, FEM module residual stress prediction: Based on the optimized heat source model, the FEM module is used to simulate the three-dimensional temperature field during welding, calculate the residual stress distribution of the welded joint, and output the stress concentration factor and deformation prediction results of the key area; S4. Multi-physics field collaborative optimization and parameter adjustment: Based on the CFD thermal flow field distribution and FEM residual stress prediction results, the welding process parameters are adjusted to optimize the heat source model to ensure that the thermal-mechanical coupling performance meets the design requirements; S5, experimental verification and model correction: prepare actual welding samples, measure the residual stress distribution by blind hole method or X-ray diffraction method, compare the FEM prediction results of step S3, and iteratively correct the simulation model according to the experimental data; S6. Process plan output and application: Integrate the optimized heat source model parameters and FEM residual stress distribution map to form a welding process guidance plan, clarify key indicators such as heat input control range and residual stress safety threshold, and directly apply it to the thick plate welding process design and quality assessment of offshore structures.

3. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 2 is characterized in that: In the CFD of step S2, the flow of the welding molten pool follows the three conservation laws of fluid motion: the law of conservation of mass, the law of conservation of momentum and the law of conservation of energy, and their expressions are as follows: （1） ； （2）； （3）； In the formula, , , , , , , , , , They are material density, time, velocity, mass source term, static pressure, viscosity, momentum source term, material enthalpy, thermal conductivity, and energy source term. The material enthalpy expression is as follows: （4）； （5）； In the formula, , , , , , , They are reference material enthalpy, reference temperature, specific heat capacity, latent heat, liquid volume fraction, solidus temperature, and liquidus temperature; In the CFD of step S2, the mushy zone is regarded as a porous medium to avoid the complex calculation of directly tracking the interface between phases. The momentum source term expression is as follows: （6）； In the formula, is the paste parameter, is the drag speed, is 0.0001.

4. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 3 is characterized in that: In the CFD of step S2, the surface tension is negatively correlated with the temperature, and its expression is as follows: （7）； In the formula, is the surface tension gradient, is the melting point, is the surface tension at the melting point.

5. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 4 is characterized in that: In the CFD of step S2, the buoyancy expression generated by the welding pool is as follows: （8）； In formula (8), is the coefficient of thermal expansion; The VOF value of the free surface of the molten pool is tracked at each time step by the fluid volume method, and its expression is as follows: （9）； In formula (9), represents the fluid volume fraction; In the welding process, in addition to heat input, the effects of convection heat transfer and radiation heat transfer on heat flux must also be considered comprehensively. The expression is as follows: （10）； In formula (10), is the heat flux density, is the convection heat coefficient, is the material emissivity, is the Stefan-Boltzmann constant.

6. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 5 is characterized in that: In the CFD of step S2, the heat source model adopts a semi-elliptical volume heat source form, which can effectively approximate the shape and size of the welding pool and accurately simulate the distribution of the mobile heat source during the welding process. The heat source expressions of the front and rear are as follows: （11）； （12）； In the formula, , , is the local coordinate system of the heat source model, , , , is the shape parameter of the heat source model, , , For welding efficiency, voltage and current, , is the front and backup heat fraction of the mobile heat source.

7. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 6 is characterized in that: In the FEM module of step S3, the material behavior in the plastic zone satisfies the plastic flow and strain hardening laws, and the total strain is divided into the following three components: (13) In the formula, is the elastic strain, is the plastic strain, For thermal strain.

8. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 7 is characterized in that: In CFD, the weld geometry is determined by the VOF method, where the function f represents the volume fraction of the metal term; In this module, the following three cases are considered: (1) when f>0.5, it represents the metal item; (2) when f=0.5, it represents the interface between the metal item and the gas item; (3) when f<0.5, it represents the gas item.

9. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 8 is characterized in that: In CFD, the temperature at the center of the Euler control volume in CFD is mapped to the Lagrangian unit node of the FEM module through three-dimensional interpolation, so as to realize the loading of temperature distribution. The expression is as follows: （14）； In the formula, , , is the local coordinate system of the CFD model, , , are the natural coordinates of the CV centers around the FEM model nodes, is the temperature value of the 8 CV centers adjacent to the unit node, is the interpolation function, is the temperature mapped to the FEM model.

10. The method for numerical prediction of residual stress in thick plate welding based on thermal-fluid-solid multi-field coupling according to claim 9, characterized in that: The two adjacent moments in the FEM model are obtained by formula (14): , Temperature and Afterwards, when Temperature between Approximate temperature by linear interpolation The expression is as follows: （15）； In order to reasonably and accurately apply the temperature distribution data provided by the CFD model, the time step of the FEM model should not be larger than the time interval of the CFD model, so as to achieve accurate interpolation between adjacent temperature distributions.

Images