Research On Variable-fidelity Metamodeling Assisted Laser Penetration Welding Process Parameters Optimization

Laser penetration welding(LPW)has been widely used in a variety of industrial applications due to its significant advantages including but not limited to precise welding energy control,high productivity,deep penetration,and high efficiency.Process parameters play a decisive role in the welding profile and the quality of welding joint.An unsuitable selection of process parameters will lead to welding defects,e.g.,hump,porosity,incomplete fusion,collapse and dripping,which can significantly decrease the mechanical and functional properties of welding joints.Therefore,the process parameters optimization is very important for improving the quality of laser penetration welding joint.However,the high non-linear relationship between the process parameters and the welding results cannot be expressed explicitly.So,the metamodels are employed to fitting the relationships between the process parameters and welding results for the sake of welding design optimization.The existing optimization methods can be divided into two distinct types: physical experiment-based optimization approach and simulation-based optimization approach.But,neither method can balance the contradiction between cost and precision in the optimization process.This issue has limited their capability of guiding the actual welding processing and improving welding quality.This paper proposes a variable-fidelity(VF)metamodel based laser penetration welding process parameter optimization approach,in which two levels fidelity information,both from low fidelity computer simulations and high-fidelity physical experiments can be integrated and fully exploited.In the proposed approach,the three-dimensional thermal finite models are developed to obtain the low fidelity data of welding results.Then,a VF metamodel is constructed to approximate the relationships between the process parameters(laser power,welding speed and laser focal position)and welding results(bead width,depth of penetration and angular deformation),and a randomized trial is employ to verify the prediction accuracy of the VF metamodel.Finally,the multi-objective optimization is used to facilitate LPW process parameter space exploration and multi-objective Pareto optima search.The verification experiments verify the effectiveness and reliability of the obtained optimal process parameters.