584 results about "Finite element calculations" patented technology

The invention discloses an urban inland inundation situation deduction method based on fusion of mechanism knowledge and deep learning, and the method comprises the steps: collecting data, such as weather, terrain and historical disaster situations, and constructing a standardized database in combination with an inland inundation production and convergence mechanism; constructing an urban inland inundation numerical twinborn model, and optimizing hyper-parameters to improve prediction precision by adopting finite element calculation and a multi-scale coupling method based on hydrological and hydrodynamic models; constructing a causal association network by using the numerical twinborn data, and identifying a core affected area; designing a situation deduction algorithm of a space-time diagram neural network, capturing rainfall waterlogging dynamic characteristics, and realizing advanced early warning and accurate deduction; based on the waterlogging intensity and the exposure and vulnerability of the disaster-bearing body, a risk assessment system is constructed, the disaster diffusion trend and the disaster risk are analyzed, and support is provided for flood prevention decision and emergency management.

The invention relates to the technical field of data processing, and discloses a comprehensive evaluation method and system for the health degree of a distribution network cable. The method comprises the following steps: collecting cable joint multi-source monitoring signals, normalizing the monitoring signals to obtain a degradation degree feature vector, correcting multi-physics field coupling model parameters, obtaining a recessive degradation index through finite element calculation to obtain an enhanced feature vector, and performing time-frequency domain decomposition to extract multi-scale feature parameters to obtain a comprehensive feature matrix; a double attention mechanism calculates a feature weight and a time sequence correlation degree to obtain a deterioration trend prediction value, and fuzzy integral is fused with a multi-classifier output probability to obtain a health degree evaluation grade and an early warning result. According to the invention, the early defect identification accuracy and the degradation trend prediction precision are improved.

The present invention discloses a method and apparatus for simulating thermal stress of a casting mold during a service process, and a storage medium. The method comprises: importing a pre-built three-dimensional geometric model of the casting mold, processing the geometric model, and then performing grid division to obtain a casting simulation finite element physical model; assigning material parameters, interface parameters, process parameters, and boundary conditions to the finite element physical model to obtain a casting simulation finite element calculation model; performing casting process simulation calculations to obtain casting simulation results such as mold temperature field and stress field, and exporting result data; interpolating a temperature-stress correction factor; multiplying all correction factor data under the same node numbers with mold stress data to obtain final thermal stress distribution data of the mold during the service process.

The invention relates to the technical field of hydraulic engineering and geotechnical engineering, in particular to a parameter updating method and device for a rock-fill dam deformation analysis model.The method comprises the steps that a plurality of Duncan-Zhang E-B model parameter combinations of an indoor triaxial experiment of a target rock-fill dam project are obtained to construct a copula multivariate joint distribution model; sampling a parameter combination sample of the target rock-fill dam project, and inputting the parameter combination sample into the finite element deformation analysis model to generate a finite element calculation sample; training a pre-constructed deep neural network by using the parameter combination sample and the finite element calculation sample to generate a deep learning agent model; and fusing the copula multivariate joint distribution model, the deep learning agent model and an NSGA-III multi-objective optimization algorithm to search for final model parameters for updating the rock-fill dam deformation analysis model generation by generation. Therefore, the problem that an existing dam project parameter updating method neglects complex correlation existing between different parameters is solved.

The present application relates to a BIM-based hanging basket design method, computer device and computer-readable storage medium, wherein the design method comprises following steps: a parameter acquisition step; a hanging basket modularization step: decomposing structures of hanging baskets into a plurality of design modules; a selection step; a calculation step: calculating parameters of the design modules using a structural mechanics calculation method and / or a finite element calculation method; preliminary BIM model building step: building models of the design modules and a preliminary BIM model of the hanging basket; an optimization iteration step: converting the preliminary BIM model of the hanging basket into a finite element model for calculation, and feeding calculation results back to the preliminary BIM model for structural adjustment, repeatedly performing iterative calculations to obtain optimized hanging basket parameter data; final BIM model acquisition step: the BIM model of the hanging basket adjusted according to the optimized hanging basket parameter data is a final BIM model of the hanging basket. The design method enables autonomous calculation and design of the hanging basket with good versatility.

The invention provides an arch dam shape efficient intelligent optimization method and system based on an agent model. The method comprises the steps that an arch dam physical-numerical model is constructed, a double optimization target with structural safety and economical efficiency as the core is determined, design parameters are selected, a constraint function is set, and an evaluation index system is established; samples are generated through Latin hypercube sampling, a data set is constructed in combination with finite element calculation, and a multi-task learning architecture is adopted to train a high-precision agent model. And then, coupling the proxy model with a multi-objective optimization algorithm, quickly searching a Pareto optimal solution set, and screening out a comprehensive optimal figure by using a multi-attribute decision-making method. And finally, through a finite element simulation verification result, prediction precision and performance improvement are ensured. The method has the advantages of lightweight modeling, efficient prediction and accurate search, and provides an effective tool for intelligent optimization and rapid decision making of hydraulic structures such as arch dams and the like.

The invention discloses a method for researching diffusion stress of a lithium ion battery by constructing a high-fidelity microstructure model. The method comprises the following steps: firstly, performing tomography on an electrode material by using a synchrotron radiation X-ray tomography microscope, identifying active particles in the electrode material, calculating geometric parameters such as spatial distribution and volume radius of the particles, and establishing a three-dimensional geometric model of the lithium ion battery; based on mathematical models such as a reaction mechanism, a kinetic equation and a mechanical equilibrium equation of the lithium ion battery, a three-dimensional heterogeneous electrochemical-force coupling model is constructed, initial and boundary conditions are set for the model, grids are divided, and finite element calculation is carried out. According to the invention, the distribution of diffusion stress and the relationship between the distribution of diffusion stress and lithium ion concentration can be analyzed, and meanwhile, the electrode characteristics such as porosity, curvature, anisotropy and uniformity of the electrode can be analyzed. And battery failure phenomena such as particle breakage and particle stripping can be avoided, and theoretical guidance can be provided for preventing battery failure and capacity fading.

The invention belongs to the technical field of bridge engineering, and particularly discloses a proxy model and finite element fusion method and system for bridge safety early warning. Comprising the following steps: simulating a real load of a bridge by adopting a structural multi-scale finite element model to calculate a bridge structural response, constructing a finite element agent model, and training the finite element agent model by adopting a data set to extract a relationship between weight matrix analysis and output items of a neural network; and in the refined grid division part, considering the damage directly observed by the structure, dividing the model parameters extracted from the monitoring point by the model parameters directly corrected by the damage to obtain a final corrected model. According to the method, sensor signal processing, model correction and finite element calculation are closely combined together through neural network analysis, a method basis is provided for dynamic evolution of a digital object fusion model, and technical support is provided for subsequent bridge structure safety evaluation and health early warning.

The invention provides a model quality evaluation optimization method and device for industrial simulation application and a server, and relates to the technical field of simulation application generalization development, and the method comprises the steps: obtaining a to-be-processed geometric model, and carrying out the format detection processing, format conversion processing and asynchronous processing of the geometric model, and determining a target input model; performing grid division processing and finite element analysis processing on the target input model, determining a finite element calculation result, performing local error estimation analysis processing on the finite element calculation result, locally refining a model grid of the target input model, determining a model key region, performing local encryption processing on the key region, and obtaining a local encryption result of the target input model; determining a model output result; and carrying out optimization processing on the model output result through a model quality evaluation optimization system with a layered architecture, and determining an optimized target model output result. The model quality can be remarkably improved.

The embodiment of the invention provides a BIM-based bridge model intelligent construction method and system, and belongs to the field of BIM-based bridge model intelligent construction. The method comprises the steps that bridge model construction related data are collected and preprocessed, and a component library is established according to the preprocessed bridge model construction related data; defining geometric linkage rules and topological connection logic of components in the component library, and generating an initial BIM bridge model through a generative adversarial network; performing lightweight finite element calculation on the initial BIM bridge model to obtain an optimized BIM bridge model; and performing standard conflict detection and conflict parameter adjustment on the optimized BIM bridge model to obtain a final bridge model. According to the method, the parameter relation matrix and the dynamic linkage mechanism are constructed, the generative adversarial network is introduced to generate the complex bridge type, and the graph convolutional network is introduced into the generative adversarial network, so that reasonable bridge type topology and dynamic parameter linkage are guaranteed.

The invention relates to the technical field of infrared light machine system thermal deformation prediction, in particular to a multi-material structure thermally induced stress deformation prediction method based on a graph neural network. The method comprises the steps of data set establishment, graph structure establishment, graph neural network model establishment and training and model and parameter optimization. Finite element nodes correspond to graph nodes, finite element edges correspond to graph edges, an encoder-message passing-decoder architecture model is established, and node states are updated through a three-layer physical symmetry message passing mechanism. Physical constraint loss including minimum displacement smoothness constraint and stress continuity constraint is innovatively added into a loss function. Compared with traditional finite element calculation, the method has the advantages that the speed is increased by more than 100 times, high hardware adaptability is achieved, the black box limitation of a data-driven neural network model is broken through, thermally induced stress deformation analysis caused by different material coefficients can be processed, the adaptability to geometric changes is high, and good engineering application value is achieved.

Disclosed in the present invention is a three-dimensional laser point cloud-based cable force measurement method for a stay cable. Stay cable actual measurement line shape fitting is performed by utilizing point cloud data, obtained by means of three-dimensional laser scanning, of the stay cable, and on the basis of a finite element calculation method, a line shape that is calculated by a stay cable finite element model and is the same as the actually measured line shape of the stay cable is found, thereby determining an actual cable force of the stay cable, and achieving the purpose of accurately measuring the cable force of the stay cable. According to the present invention, a three-dimensional laser scanning technology is introduced and applied to cable force measurement of a stay cable, and for the defects of a common cable force measurement method, the characteristics of non-contact measurement, high measurement precision, high measurement efficiency and the like of the three-dimensional laser scanning technology are utilized so as to achieve rapid measurement of the line shape and the section size of the stay cable. On the basis of finite element calculation and analysis, the three-dimensional laser point cloud-based cable force measurement method for the stay cable is provided, thereby achieving rapid and efficient measurement of the cable force of the stay cable.

The invention discloses a simulation optimization method, device and system for tire multi-rubber modulus matching, and belongs to the technical field of tire simulation optimizing.The method comprises the steps that a tire and a pavement finite element model are built, and the tire and pavement finite element model comprises a three-dimensional tire simulation model and a two-dimensional rigid pavement model; matching a modulus attribute of a rubber material to which each finite element grid of the tire simulation model belongs through an orthogonal test design method, carrying out finite element calculation based on the tire and the pavement finite element model, and calculating multiple groups of stress-strain and strain energy data of target belted layer endpoints of the tire under different modulus attributes; and respectively calculating modulus characteristic values of the tire under different modulus attributes on the basis of the multiple groups of stress-strain and strain energy data of the endpoints of the target belted layer, so as to obtain the optimal modulus ratio of multiple rubber materials of the tire. According to the method, the precision and usability of simulation optimization are improved, the service life of the tire is prolonged, the utilization rate of tire design materials is increased, and the design cost is reduced.

The invention discloses a numerical simulation method for simulating resin flow in a molding process of an intelligent composite material with an embedded optical fiber. The method comprises the following steps: generating a fiber bundle model through a Monte Carlo random algorithm; establishing a mesoscopic model for embedding optical fibers inside or outside the fiber bundle; establishing a resin flow control equation based on a phase field method in the finite element model; applying pressure load to an inlet and an outlet of the model, and submitting finite element calculation; the velocity field and the pressure field obtained through simulation are checked, and the influence rule of the optical fiber on the resin infiltration fiber bundle is analyzed. The method has good guiding significance for optimizing the intelligent composite material forming process and reducing forming defects.

The invention belongs to the field of winding temperature control, and relates to a PCA-FCNN-based converter transformer valve side winding oil flow temperature rise calculation method, which analyzes a thermodynamic conduction process in a valve side winding, determines a characteristic variable having the highest correlation with a temperature field, performs high-correlation characteristic variable parameterization setting by using Latin hypercube sampling, and calculates the oil flow temperature rise of the valve side winding. A valve side winding flow-heat coupling calculation model is established, and finite element calculation software is used for flow-heat coupling calculation; constructing a data set of a calculation model by taking the parameterized characteristic variable as an input characteristic quantity and taking the valve side winding oil flow temperature rise calculated by a finite element as an output characteristic quantity, dividing the data set into a training set and a test set, and generating a valve side winding oil flow temperature rise rapid calculation model with relatively high precision by setting model training parameters; a validation set other than the model data set is set to validate the accuracy of the computational model. According to the method, the required calculation time is shortened by about 300 times compared with the calculation time of a finite element method, and the calculation speed of the oil flow temperature rise of the valve side winding is increased.

The invention discloses an old transformer hot-spot temperature prediction method based on multi-source heterogeneous data processing. The method comprises the following steps: step 1, establishing a transformer digital twinborn analysis model; 2, taking the transformer physical model as an object, obtaining optical fiber temperature measurement data, thermocouple temperature measurement data and an infrared temperature measurement picture, and proposing a heterogeneous data processing technology to carry out data extraction on the infrared temperature measurement picture; 3, finite element calculation is carried out through magnetic-thermal-fluid multi-physics field coupling, and correction is carried out through experimental data obtained in the step 2; 4, temperature data is obtained according to finite element calculation in the step 3 to serve as a data analysis basis, and then a random forest model is established for training and optimization; step 5, obtaining operation temperature data of the transformer; and step 6, performing prediction calculation on the hot spot temperature. The method has the advantages that compared with traditional hot spot temperature calculation in which a temperature field is simplified, the method can effectively avoid the inaccurate situation.

The invention provides an in-situ real-time monitoring method and system for a rock microscopic fracture mechanism and stress evolution. The method comprises the following steps: S1, standardized preparation and pretreatment of a rock sample; s2, in-situ loading and multi-modal data synchronous acquisition are carried out; s3, crack dynamic tracking and stress-strain field joint calculation; s4, constructing a microscopic-macroscopic quantitative correlation and damage model; and S5, performing multi-dimensional dynamic visualization and engineering output. The in-situ scanning electron microscope and the mechanical loading device are integrated, so that cross-scale synchronous monitoring and quantitative analysis of micro crack evolution and macroscopic stress-strain response are realized. A low-disturbance clamping device and a conductive anti-charge sample are treated, and in-situ testing is ensured; combining DIC and deep learning to quantify a crack propagation path, a local strain field and energy characteristics; through DIC strain inversion and finite element calculation, a non-contact stress cloud picture is generated in real time, and finally quantitative correlation between a microscopic fracture mechanism and macromechanical response is established.

The invention provides a pavement structure response monitoring and total distribution calculation method based on a digital twinborn technology. The method comprises the following steps: S1, constructing a road structure and a digital twinborn model thereof; s2, an embedded sensor is arranged in the road structure, and real-time structure response data of the road structure at the feature position under the vehicle load effect are obtained; s3, carrying out an on-site loading test, obtaining structure response actual measurement data under the action of a vehicle moving load, and verifying and correcting the digital twinborn model; s4, using finite element calculation software to perform simulation calculation and output structural response data of the sampling points under the action of different load working conditions based on the digital twinborn model; and S5, training a deep learning model by using the simulation response data. According to the method, a digital twinning technology is utilized, a high-fidelity mapping model of a physical entity in a virtual space is constructed, behaviors of the physical entity in a real environment are accurately simulated, and a data basis can be provided for a machine learning response deduction method at low cost.

The invention discloses a containment failure probability curve fitting method and system based on a random forest, and relates to the technical field of nuclear power safety. Comprising the following steps: acquiring key parameters influencing ultimate bearing capacity of a containment, determining a distribution interval of the key parameters, and establishing a key parameter table; performing finite element calculation on the basis of the key parameter table, obtaining ultimate bearing capacity of the containment vessel under different parameter combinations and temperature loads, and establishing a database; performing normalization preprocessing on the database to form a training data set; establishing a random forest model and performing regression model training by using the training data set; inputting actual design parameters of the containment and the temperature required to be evaluated into the trained model to obtain an ultimate bearing capacity predicted value of the containment; and fitting of a failure probability curve is carried out based on the central limit theorem. According to the method, complex probabilistic safety analysis is converted into a standardized calculation task, and the implementation threshold of nuclear power plant design verification and safety review is greatly reduced.

The invention relates to the technical field of electrics, in particular to a cable core temperature prediction method, system and device and a medium, and the method comprises the steps: building a multi-physical field coupling finite element model of a cable based on a first target parameter and a target condition; performing finite element calculation on the finite element model to generate a sample data set under multiple working conditions; performing preset processing on the sample data set to generate multi-dimensional time sequence input data; training a deep learning model by using the multi-dimensional time sequence input data to obtain a target model; and performing cable core temperature prediction on the acquired cable operation parameters by using the target model. The beneficial effects are that intelligent temperature monitoring and early warning can be provided for safe operation of the cable, an overheating fault is prevented in advance, stable power supply of a power grid is guaranteed, and the operation and maintenance cost and risk are reduced.

The invention relates to the field of dam effect quantity monitoring, in particular to a dam monitoring effect quantity prediction method based on mechanism and data dual drive, which comprises the following steps of: performing noise reduction preprocessing on monitoring data by adopting a wavelet noise reduction method; constructing a monitoring effect quantity long-term prediction model; constructing a short-time proximity precise prediction model; a Newton-Raphson optimization algorithm (NRBO) is used for optimizing network hyper-parameters of a physical constraint radial basis function (PIRBF) to construct an inversion agent model of finite element model parameters, a finite element calculation result is calibrated, and finally a hybrid model is constructed to realize long-term prediction of monitoring effect quantity. A deep learning model is constructed through an aurora optimization algorithm (PLO), a Transform architecture and a gated cycle unit network (GRU) to correct a long-term prediction model error, a deep learning model result is coupled to construct a short-term approaching prediction model, and short-term approaching accurate prediction of the effect quantity is realized.

The invention provides a hysteresis loss calculation method suitable for motor electromagnetic field finite element calculation, and relates to the technical field of motor simulation calculation. The method comprises the following steps: calculating flux density vector distribution of a motor core grid through a transient electromagnetic field finite element; calculating a fundamental wave period according to the vector distribution; selecting a flux density vector in a fundamental wave period, performing flux density ellipse fitting, and rotating an ellipse along an original point to enable a long axis of the ellipse to be parallel to a Y axis; translating the long axis of the ellipse along the X axis, and enabling the long axis to coincide with the Y axis; establishing a hysteresis loss calculation model; checking parameters through a hysteresis loss curve, and calculating to obtain Kx; and calculating the magnetic hysteresis loss of all grids, and summing to obtain the magnetic hysteresis loss of the whole motor iron core. By analyzing the flux density of each grid of the iron core and combining the pulsating magnetic hysteresis loss curve of the silicon steel sheet, a motor magnetic hysteresis loss algorithm model considering rotation and pulsating loss is obtained, and the precision of finite element calculation of the motor magnetic hysteresis loss is improved.

The invention belongs to the technical field of machining, and particularly relates to a thin-wall structure milling boss clamping layout optimization method considering a feeding path. The method comprises the steps that multiple boss clamping layout schemes are obtained; constructing a thin-wall blank finite element model, and performing finite element calculation to obtain initial residual stress data; building a thin-wall blank milling boss clamping finite element model according to different boss clamping layout schemes, performing grid division, and applying the initial residual stress data to the thin-wall blank milling boss clamping finite element model; a milling feed path is selected, milling force is gradually applied to a milled unit of the thin-wall blank milling boss clamping finite element model according to the advancing distance of a milling cutter, the milling process is simulated, and machining deformation data and residual stress data are obtained; and determining optimal technical indexes of the boss clamping layout scheme, and screening out the boss clamping layout scheme meeting the requirements according to the technical indexes, the machining deformation data and the residual stress data.

The present invention belongs to the technical field of metal plastic processing, and is specifically a method for predicting springback in roll bending of aluminum alloy profiles taking into account anisotropy. The method performs theoretical analysis based on the Mises and Hill48 yield criteria to obtain a formula for the curvature before and after springback. The roll bending data after springback compensation obtained through theoretical analysis is input into the model, and the profile equal curvature rolling analysis model is calculated. An edge line of the reconstructed profile digital model is extracted as a characteristic curve for curvature analysis. The present invention makes full use of the finite element calculation and analysis process, and uses ABAQUS to import the anisotropic parameters of the aluminum alloy into the model for numerical simulation. The theoretical analysis is verified by the simulation results, thereby achieving the purpose of reducing springback during the profile bending process.

The invention discloses a multi-billet continuous casting billet rough rolling pass system optimization method based on finite element software, and relates to the technical field of steel rolling casting, and the method comprises the following steps: based on the principle that metal second flow is equal, establishing a digital model of volume flow dynamic balance during rolling of billets with different sections, and quantifying a billet rough rolling deformation rule; carrying out hole pattern modeling of multi-blank pattern switching by utilizing the finite element simulation model, building a cooperative computing framework, and dynamically deploying a finite element computing scheduling task; on the basis of multi-physics coupling analysis, stress-strain characteristics and size evolution rules of the blanks under the condition of multi-dimensional parameters are predicted, a process optimization target is generated, and the requirements of the blanks with different sections are met; and real-time mapping of virtual and real data is established, a process knowledge base is constructed, and regular learning and updating are carried out. According to the method, the finite element software is adopted, so that the rough rolling deformation rule of different section blank types can be simulated more accurately, the shape and the size of a rolled material can be predicted and controlled more accurately, and the accuracy and the reliability of an optimization result are improved.

The invention belongs to the technical field of airplane structure strength digital twinborn model construction, and particularly relates to an airplane structure finite element calculation result data noise reduction method, which comprises the following steps: step 1, preliminarily identifying deformity units including triangular units and high slenderness ratio units; identifying the triangular units based on the continuity of the node numbers; identifying a high slenderness ratio unit based on the side length; 2, defining a stress concentration area, and removing units in the stress concentration area from the deformed units; defining a stress concentration region including a hole edge, a fillet, a boundary constraint position and a region where a concentrated load applying position is located; 3, calculating the gradient of the deformity units, and removing units in a real physical gradient field from the deformity units; step 4, carrying out noise reduction processing on the calculated value of the malformation unit; and smoothing the calculated value of the malformed unit, or removing the calculated value of the malformed unit.

The invention relates to a site micro-vibration attenuation test measuring point determination method and system, belongs to the technical field of vibration test, and solves the problem that vibration test measuring point arrangement depends on vibration source load accuracy or artificial experience in the prior art. The method comprises the following steps: determining a basic measuring point based on a target site parameter, obtaining an actual measurement vibration response, and calculating an actual measurement response transfer ratio; establishing an initial finite element model of the target site, and applying a preset load to the initial finite element model to obtain an initial finite element calculation result; according to the basic measuring points, corresponding model calculation points and predicted vibration responses are extracted, and predicted response transfer ratios are calculated; and performing model parameter updating by using the actually measured response transfer ratio and the predicted response transfer ratio, obtaining a basic finite element model, generating a vibration transfer form curve, obtaining vibration transfer characteristics, and determining the number and the positions of the measuring points. The micro-vibration transmission characteristics are accurately determined without depending on the vibration source load, and the number and the positions of the vibration attenuation test points are reasonably formulated.

The invention provides an automatic pretreatment format conversion method for parachute inflation finite element simulation. The automatic pretreatment format conversion method comprises the following steps: carrying out geometric modeling and finite element mesh division on a canopy and a parachute cord by utilizing Gmsh; a preprocessing format conversion system is developed, grid files and calculation parameters are uploaded, and the conversion system automatically stores, calculates and integrates data and converts preprocessing files into input files needed by finite element calculation; then, finite element simulation analysis of the typical parachute inflation process is carried out through CalculiX, so that the reliability of the automatic conversion function of the pretreatment format is verified; the automatic preprocessing format conversion system suitable for parachute inflation finite element simulation and the graphical user interface of the automatic preprocessing format conversion system are developed, full-automatic processing of files is achieved, the use threshold of a user is lowered, and simulation efficiency and application benefits are improved.

The invention discloses a concrete panel tensile crack damage calculation method and a concrete panel tensile crack damage calculation system, which combine extended finite element calculation with a concrete tensile crack damage calculation model, accurately simulate crack behaviors in concrete, and are particularly suitable for numerical simulation requirements under the conditions of complex crack networks and large-scale damage. According to the method, the expansion finite element method is introduced, subtle changes of crack starting, expansion paths and final forms are captured on the premise that the calculation cost is not increased, the common problem of grid dependence in a traditional finite element method is avoided, fine division or frequent updating of grids is not needed, and the calculation efficiency is improved. The method reduces the preprocessing time and iteration times, reduces the consumption of overall computing resources, can accurately describe the expansion path of the crack and the nonlinear mechanical behavior of the concrete, and is suitable for the simulation of a complex crack network.

The invention discloses a welding simulation heat source parameter calibration method and system based on a hybrid agent model, and belongs to the technical field of welding numerical simulation and optimization, and the method comprises the following steps: S1, problem definition and experiment design; s2, analyzing a product structure and constructing a joint model; s3, parameter space sampling; s4, performing simulation calculation; s5, constructing a mixed agent model; s6, proxy model optimization based on a genetic algorithm; s7, parameter verification and model updating; and S8, optimizing and outputting optimal parameters. According to the method, most time-consuming full-order finite element calculation is replaced by the proxy model and the simplified joint model, the parameter calibration process is shortened from several days or weeks to several hours, and the optimization efficiency is improved; a simplified model and a complete model are used for optimization, firstly, the simplified model is used for rapidly obtaining an optimal parameter approximate value, and the complete model is used for obtaining accurate parameters.