一种冷丝填充速度的GABP优化算法

A GABP optimized algorithm for filler rate of non-heated wire

冷丝填充埋弧焊过程中,冷丝填充量是决定焊缝组织和性能的主要参数.通过大量的工艺试验研究了冷丝填充量对微观组织和力学性能的影响,利用焊接电弧热平衡规律建立了冷丝填充过程热量动态分配平衡方程,推导出了计算冷丝填充速度的关系式.采用人工神经网络实现了焊接电流、电弧电压、焊接速度与冷丝填充速度的非线性映射关系.结果表明,基于遗传算法的BP神经网络(BP neural network based on genet-ic algorithm,GABP)优化算法实现了冷丝填充埋弧焊过程的自适应控制,实际焊接冷丝填充速度与期望值之间的线性相关度达到0.991 88,表明该算法可以满足冷丝填充埋弧焊工艺及性能要求.

For the process of submerged arc welding filled with non-heated wire,the filler quantity of non-heated wire is one of the main parameters that affects the microstructure and mechanical properties of welded joints.First,through a lot of welding experiment,the effect of the filler quantity of non-heated wire on the microstructure and mechanical properties was investigated.On the basis of the heat balance law in welding process,the balance equation of heat dynamic distribution for the process of submerged arc welding filled with non-heated wire was established.Then the relational expression was formulated for the filler rate of non-heated wire.At last,the nonlinear mapping relationship between the welding current I,arc voltage U,welding speed v and filler rate of non-heated wire vl was realized based on artificial neural network.Experimental results showed that,the optimized algorithm of BP neural network based on genetic algorithm(GABP) can realize adaptive control for the process of submerged arc welding filled with non-heated wire.The linear correlation between filler rate of non-heated wire in actual welding process and the expectative output comes up to 0.991 88.This shows that the algorithm of GABP can meet requirement for welding process and properties.