Welding Sequence Optimization Research Articles

Parameter analysis and optimization of welding sequence for the aluminium alloy sidewall in a high-speed train body

The sidewall is the most complex structure and welding process part in the aluminum alloy train body, and its welding deformation is greatly influenced by welding sequence parameters. In this paper, the relationship between the welding residual stress and welding deformation of the sidewall and welding sequence parameters is studied, and the optimal welding sequence is determined. Using a verified finite element model of one intercepted sidewall, the relationships between the welding residual stress and welding deformation of the sidewall and the welding of welds, welding sequences of welds and the weld sequence are studied, respectively. Some conclusions can be drawn from the analysis results. One is that the welding results at one weld are only determined by self-welding, and have nothing to do with other welds, and welding results outside one weld are affected by welds on both sides. The second is that for a single weld, the from-one-end-to-the-other welding sequence is best, and the identical welding sequence is best for each weld. The third is that the fixed welds are firstly welded and a cross weld sequence is best.

Rapid sequence optimization of spot welds for improved geometrical quality using a novel stepwise algorithm

ABSTRACT Joining sequence optimization is a combinatorial problem, requiring extensive computational time. The significance of determination of an optimal sequence for improved geometrical quality is substantial. Previously, genetic algorithms have been studied for defining the optimal sequence. However, these algorithms are highly dependent on the internal parameters, requiring additional computational analysis and thereby extended evaluation time. In this article, a novel robust stepwise algorithm is introduced to determine the optimal weld sequence. Application of the proposed algorithm leads to drastic time improvements for defining the optimal weld sequence of each assembly. Three industrial assemblies are evaluated. Comparison with the previously applied population-based optimization algorithms indicates that the optimization time can be reduced drastically with the proposed stepwise algorithm. The stepwise algorithm is intended to be applied in a geometry assurance digital twin, where the assembly parameters are being optimized for each individual assembly.

Efficient Spot Welding Sequence Optimization in a Geometry Assurance Digital Twin

Abstract A digital twin for geometry assurance contains a set of analyses that are performed to steer the real production for securing the geometry of the final assembly. In sheet metal assemblies, spot welding is performed to join the parts together. The sequence of the welding has a considerable influence on the geometrical outcome of the final assembly. In industry, the sequence of welding to secure the geometry is mainly derived by tacit manufacturing knowledge. Including such knowledge to mimic the production process requires extensive knowledge management, and the result might be just a good enough solution. Theoretically, spot welding sequence optimization for the optimal geometrical quality is among NP-hard combinatorial problems. In a geometry assurance digital twin, where assembly parameters are selected for the individual assemblies, time constraints define the quality of the optimal sequence. In this paper, an efficient method for spot welding sequence optimization with regards to the geometrical quality is introduced. The results indicate that the proposed method reduces 60–80% of the time for the sequencing of the spot welding process to achieve the optimal geometrical quality.

Incorporating domain knowledge into reinforcement learning to expedite welding sequence optimization

Welding Sequence Optimization (WSO) is very effective to minimize the structural deformation, however selecting proper welding sequence leads to a combinatorial optimization problem. State-of-the-art algorithms could take more than one week to compute the best sequence for an assembly of eight weld beads which is unrealistic for the early stages of Product Delivery Process (PDP). In this article, we develop and implement a novel Reinforcement Q-learning algorithm for WSO where structural deformation is used to compute reward function. We utilize a thermo-mechanical Finite Element Analysis (FEA) to predict deformation. The exploration–exploitation dilemma has been tackled by domain knowledge driven ε-greedy algorithm into Q-RL which helps to expedite the WSO and we call this novel algorithm as DKQRL. We run welding simulation experiment using well-known Simufact® software on a typical widely used mounting bracket which contains eight welding beads. DKQRL allows the reduction of structural deformation up to ∼71% and it substantially speeds up the computational time over Modified Lowest Cost Search (MLCS), Genetic Algorithm (GA), exhaustive search, and standard RL algorithm. Results of welding simulation demonstrate a reasonable agreement with real experiment in terms of structural deformation.

A new surrogate model\u2013based method for individualized spot welding sequence optimization with respect to geometrical quality

In an individualized shee metal assembly line, form and dimensional variation of the in-going parts and different disturbances from the assembly process result in the final geometrical deviations. Securing the final geometrical requirements in the sheet metal assemblies is of importance for achieving aesthetic and functional quality. Spot welding sequence is one of the influential contributors to the final geometrical deviation. Evaluating spot welding sequences to retrieve lower geometrical deviations is computationally expensive. In a geometry assurance digital twin, where assembly parameters are set to reach an optimal geometrical outcome, a limited time is available for performing this computation. Building a surrogate model based on the physical experiment data for each assembly is time-consuming. Performing heuristic search algorithms, together with the FEM simulation, requires extensive evaluations times. In this paper, a neural network approach is introduced for building surrogate models of the individual assemblies. The surrogate model builds the relationship between the spot welding sequence and geometrical deviation. The approach results in a drastic reduction in evaluation time, up to 90%, compared to the genetic algorithm, while reaching a geometrical deviation with marginal error from the global optimum after welding in a sequence.

Method to systemically order welding sequence to efficiently mitigate welding displacement of a general ship grillage structure

ABSTRACT Fabrication of steel materials using high heat processes, such as welding, for building structures may lead to unexpected complex distortion. Thus, the optimisation of the assembly sequence to obtain the lowest deformation for huge steel structures, such as ships and offshore structures, is crucial. The objective of this study is to introduce an efficient method to systemically determine the optimal welding sequence for the lowest deformation of a general ship side panel, which is widely employed to design vessels and offshore structures. In this study, numerical simulation with a finite element method based on the inherent strain, interface element and multipoint constraint function is used as a precise computational approach to analyse the welding deformation. The employed numerical simulation obviously validated proposed systemic method to efficiently decide the optimal welding sequence for minimising welding displacement.

A Novel Rule-Based Method for Individualized Spot Welding Sequence Optimization With Respect to Geometrical Quality

Abstract Spot welding is the predominant joining process for the sheet metal assemblies. The assemblies, during this process, are mainly bent and deformed. These deformations, along with the single part variations, are the primary sources of the aesthetic and functional geometrical problems in an assembly. The sequence of welding has a considerable effect on the geometrical variation of the final assembly. Finding the optimal weld sequence for the geometrical quality can be categorized as a combinatorial Hamiltonian graph search problem. Exhaustive search to find the optimum, using the finite element method simulations in the computer-aided tolerancing tools, is a time-consuming and thereby infeasible task. Applying the genetic algorithm to this problem can considerably reduce the search time, but finding the global optimum is not guaranteed, and still, a large number of sequences need to be evaluated. The effectiveness of these types of algorithms is dependent on the quality of the initial solutions. Previous studies have attempted to solve this problem by random initiation of the population in the genetic algorithm. In this paper, a rule-based approach for initiating the genetic algorithm for spot weld sequencing is introduced. The optimization approach is applied to three automotive sheet metal assemblies for evaluation. The results show that the proposed method improves the computation time and effectiveness of the genetic algorithm.

Effect of Welding Sequence on Welding Deformation and Re-sidual Stress of Ear Plate Structure

In this study, a three-dimensional finite element model of the welding process of the ear plate structure was established based on the thermal elastic-plastic method. The DOE method was used to design the welding sequence optimization scheme, and the effect of welding sequence on welding deformation and residual stress was studied. The results show that the variation of welding sequence has little effect on the residual stress of the ear plate structure, but has great influence on the post-weld deformation. Considering the field welding conditions, the reasonable welding sequence was determined finally, and it could meet the assembly design requirements of ear plate structure.

Investigation on welding sequence of I-beam by hybrid inversion

A thermal-mechanical model based on welding experiment and finite element (FE) method was developed for welding deformation analysis and welding sequence optimization of arc welding process used in hull steel structure forming process. By means of digital image correlation (DIC) and electrical measurement method, the deformation of T-beam was measured, and the equivalent heat source parameter (EHSP) was inversed by an optimization algorithm. The purpose of the inversion is to match the numerical results with the experiment results, so the EHSP has taken account into the influence of various factors on the welding deformation and residual stress. Thus, the developed method can avoid using the highly physical nonlinearity algorithm when considering the complex material constitutive and phase transformation process, which greatly improves the efficiency of calculation. Finally, the EHSP obtained from the hybrid inversion is applied to investigate the influence of welding sequence on welding deformation and residual stress in I-beam. This method can be applied to other metal materials and weld form.

Robust Virtual Welding Process Optimization

The finished geometry of a welded structure is greatly affected by heat-induced distortion from the welding process. This often requires correction by straightening and top-level machining to ensure manufacturing quality. Distortion is sensitive to weld sequencing and manufacturing environment. Optimal weld sequencing can significantly reduce distortion and manufacturing cost. However, an optimal weld sequence is not always intuitively obvious, especially for large structures with many welds. Virtual Fabrication Technology (VFT) is Caterpillar’s proprietary software for the simulation of welding physics via finite element analysis. Powered by the highly successful VFT platform, a combinatorial genetic algorithm based optimization approach combined with an uncertainty quantification (UQ) module for evaluating robustness has been developed at Caterpillar for weld sequence optimization (WSO). WSO has been successfully applied to more than 50 structures, resulting in the minimization of straightening operations. The combined VFT and robust WSO approach results in an accelerated manufacturing process development schedule. We will present a summary of our WSO approach and some results that demonstrate how WSO has proven to be a transformative manufacturing planning technology at Caterpillar.

Evaluating evolutionary algorithms on spot welding sequence optimization with respect to geometrical variation

Spot welding is the prevalent joining process in the automotive industry. The spot welding sequence has a notable effect on the geometrical variation of the final assembly. Finding the optimal weld sequence for geometrical quality is a fast growing and NP-complete problem. Using exhaustive search for this purpose can be a time-consuming task. In this paper, genetic, particle swarm and ant colony optimization algorithms are applied to three industrial reference cases. The performance of these algorithms for finding the optimal sequence with respect to geometrical variation is compared considering the number of function evaluations.

Deformation and Residual Stress Based Multi-Objective Genetic Algorithm for Welding Sequence Optimization

Deformation and Residual Stress Based Multi-Objective Genetic Algorithm for Welding Sequence Optimization

Numerical analysis of welding deformation and residual stress in marine propeller nozzle with hybrid laser-arc girth welds

Deformation and residual stress of the large marine propeller nozzle in hybrid laser-arc girth welding were studied. Welding distortion and residual stress of the basic T-joint were firstly analyzed by thermal-elastic-plastic finite element method and verified by experiment. Combination form of the Gaussian surface heat source model and the conical heat source model was proposed to simulate the thermal flux of laser and arc power at the weld zone. The prediction errors of the width and penetration of the weld profile were respectively 5.26% and 5.89%. For angular distortion, experiment results have a good consistent with simulation results. An obviously gradient of the longitudinal stress and transverse stress was appeared at interface zone of flange and web. An optimal welding sequence was then obtained by analyzing welding deformation and equivalent residual stress. Based on the local-global mapping method, welding deformation of the marine propeller nozzle was further predicted by loading plastic strain sourced from local model.

Assessment of HCLL-TBM optimum welding sequence scenario to minimize welding distortions

The ITER HCLL-TBM (Helium Cooled Lithium Lead Test Blanket Module) box assembly development implies the welding development of the following components: the Box and the Stiffening Grid (SG) made of vertical and horizontal Stiffening Plates (noted respectively v-SP and h-SP). This multi-chamber box structure in EUROFER97 steel is made of plates cooled by multiple meandering channels where circulates pressurized helium. For the assembly of these components, characterized by numerous multipass welds, Gas Tungsten Arc Welding (GTAW) is envisaged as reference process. Moreover, the TBM has large dimensions and thin plates which makes it very sensitive to welding distortions and is problematic regarding the assembly feasibility and compliance with geometric tolerances. This paper presents the numerical simulation and experimental work performed to optimize the v-SP to box assembly sequence, which is the most critical assembly regarding distortions, in order to minimize welding distortions.One of the technical lock of this study is high calculation times needed for this large component which implies to set up a simplified welding simulation method. The study is composed of three main phases: an experimental-numerical study of a T-joint fillet mock-up GTAW used to develop the preliminary welding procedure and to validate a simplified simulation method, a numerical optimization of the v-SP to box welding sequence via the simplified method, and the experimental application of the optimized v-SP to Box welding sequence on the TBM mock-up. The calculation and comparison of four different v-SP to box welding sequences allowed to identify the best sequence regarding welding distortions and to apply it experimentally.

Weld distortion prediction of the CFETR vacuum vessel by inherent strain theory

The CFETR (China Fusion Engineering Test Reactor) vacuum vessel sectors have high density welded joint, and as-welded size of the vacuum vessel is within a very tight tolerance, prediction of weld distortion is critical to allow the final assembly with the other components. Based on inherent strain theory, the weld distortion of the vacuum vessel can be predicted using the computer package ABAQUS. In this paper, calculation and application method of inherent strain on various welded joints is proposed. Six possible welding sequence schemes of the PS3 segment are provided within the actual processing condition, weld distortion of vacuum vessel on six welding sequence schemes is studied without external constraints condition, then according to the distortion value and distortion tendency without external constraints condition, the external constraints condition are provided. The optimal welding sequence can be obtained by the simulation result with external constraints condition, which can provide data support and theoretical guidance for the actual processing of the CFETR vacuum vessel.

Prediction of welding sequence induced thermal history and residual stresses and their effect on welding distortion

Stiffened plate panel is the major structural part of a fabrication industry where fillet welding joint is one of the most important fabrication techniques. Large stiffened structures are generally joined by several welding passes which generates thermal stresses and angular deformation. Tensile residual stresses which are generated due to welding in the weld region may lead to early failure of the structure when subjected to cyclic loading. The weld-induced residual distortion causes dimensional inaccuracy and needs rework to achieve the desired shape. Use of multiple welding passes without any optimized welding sequences typically leads to an increased degree of nonuniform heating and cooling, i.e., creating complex welding residual stress and angular deformation in the structure. In this present study, the effect of four different welding sequences on submerged arc welded fillet joint has been studied. A finite element-based numerical model has been developed to predict the thermal profile, welding residual stress, and angular deformation. The developed model considers temperature-dependent material property and material deposition by using element death and birth technique. The results have been compared with experimental one. In the effect of welding sequence on residual stress, angular deformation has been studied. Thus, the developed model presents the effect of welding sequence on the weld induced residual stresses and distortions which provide one of the most optimal welding sequence for enhanced fabrication process.

An Elitism Based Genetic Algorithm for Welding Sequence Optimization to Reduce Deformation

An Elitism Based Genetic Algorithm for Welding Sequence Optimization to Reduce Deformation

Optimization of the welding sequence and direction for the side beam of a bogie frame based on the discrete particle swarm algorithm

For any large weldment with many welding seams, the welding sequence and direction have a strong influence on the assembly and service performance, especially for the side beam of the bogie frame of a high-speed rail passenger car (CHR3; CRRC, Changchun, China). Because different combinations of the welding sequence and direction greatly increase the computational time and research costs, a three-dimensional finite element approach was developed to investigate the optimal welding sequence and direction. Then, a surrogate model was established by design of experiment and used the concepts of a pointer and stack. Finally, the welding sequence and direction were optimized by the discrete particle swarm optimization algorithm. The max residual deformation and stress of the optimal result were −3.92 mm and 212.56 MPa, respectively, which is approximately 22% and 38% lower than the traditional enterprise plan, respectively. Furthermore, a weighted form of the residual deformation and stress was proposed to the end of optimum comprehensive effect, and the result also had 11% and 28% reduction, respectively. The simulation result of the optimal plans well reproduced the theoretical distribution results of the residual deformation and stress. It is proven that the optimal result can improve the welding quality and process of the side beam weldment in production.

Blending Glowworm Swarm Optimization Algorithm with Thermal-Mechanical Coupling Finite Element Model for Optimization Method of Box Structure Welding Technology

Welding sequence is the important influence factor of the temperature field and residual stress and deformation in welding structure, At present, determining the welding sequence is usually based on experience and test method. For complex structure it is very difficult to get the optimum welding sequence. This study is aiming to settle the problem of solving the optimum welding sequence of complex box welding structure. In our study, we take the welding deformation as objective function, and then, associate glowworm swarm optimization algorithm with thermal-mechanical coupling and nonlinear thermal elastoplastic finite element model to optimize numerical simulation, eventually the optimum welding sequence is determined. Optimization results demonstrate that the optimum welding sequence obtained by our method has small welding deformation and little change rate. At the same time, our method is faster, more accurate, more effective than the traditional experience or test method.

Combining soft computing techniques and the finite element method to design and optimize complex welded products

One of the main objectives when designing welded products is to reduce strains and deformations. Strains can cause excessive angular distortion. This results in a welded product that does not meet acceptable tolerances. The geometry of the weld bead height and width depends on the input parameters speed, voltage and current, and provides the welded joint with strength and quality. As welded products become increasingly complex, deformations become more difficult to predict as they depend greatly on the welding sequence. This paper shows how a combination of the Finite Element Method, Genetic Algorithms and Regression Trees may be used to design and optimize complex welded products. Initially, Artificial Neural Networks and Regression Trees that are based on heuristic methods and evolutionary algorithms were used in predicting the weld bead geometry according to the input parameters. Then, thermo-mechanical Finite Element models were created to obtain the temperature field and the angular distortion using the weld bead geometry that the best predictive models generated. Finally, optimization techniques that are based on Genetic Algorithms were used to validate these Finite Element models against experimental results, and to subsequently find the optimal welding sequence to use in the manufacture of complex welded products.