Optimization Of SMAW Process Parameters Based On Improved GA-BP Algorithm

Welding is a key technology of modern machinery manufacturing, and theSMAW is the most common.However, the biggest flaw of SMAW is more easy toproduce serious welding deformation. In this paper, depending on the needs oftechnical department of HongChang Industry and to solve the problems of weldingdeformation existing in the company’s car dumper, combining with the company’sactual welding situation, according to the principles of practice and feasible, based onan in-depth understanding of mechanism, factors and control technology of weldingdeformation. Made use of improved GA-BP algorithm to optimize welding process toreduce welding deformation.This paper made comprehensive analysis on welding process of HongChangHeavy Industry Co. Ltd. Determining the process factors leading to weldingdeformation objectively, obtaining the direction of process optimization. Therefore,the strategy that combining welding deformation prediction which based on improvedGA-BP algorithm and optimization of multi-objective process parameters is used torealize the optimization of welding parameters.Because of the nonlinear relationship between welding process parameters andwelding deformations is very complex, so in this paper, set up a prediction model ofdeformation with using welding current, voltage, speed and cooling speed as inputsand deformation of transverse contraction and deformation of angular as outputs, toconstruct a prediction model of deformation with small sample and high precision.Then using the model to realize the relationship between process parameters ofwelding and the welding deformations, making a foundation for the multi-objectiveoptimization of process parameters.In the optimization of multi-objective process parameters, using the nonlinearrelationship which getting by the model of deformation prediction based on improvedGA-BP algorithm to replace the objective function, combined with the actual situationof welding production to select range of optimization variables, namely the weldingprocess parameters. Constructing optimization space of orthogonal principle.Usingthe deformation of transverse shrinkage and angular distortion minimum as the goalof optimization to realize the optimization of welding parameters.In addition, in this paper, designing test scientifically to demonstrate the resultsof deformation prediction model.The results of demonstrating showed that the modelis efficient and available. Moreover, making use of Matlab to develop a software ofoptimization of welding parameters, to promote the value of practice of the theory ofthis paper.