基于LPSO与BP神经网络电阻点焊工艺参数建模优化  

The Modeling and Optimization of Resistance Spot Welding Process Parameters Based on LPSO and BP Neural Network

作　　者：刘伟[1,2] 郭猛[1] 

机构地区：[1]东北石油大学电气信息工程学院,黑龙江大庆163318 [2]北京工业大学机械工程与应用电子技术学院,北京100022

出　　处：《组合机床与自动化加工技术》2016年第2期138-140,共3页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家科技重大专项资助项目(2009ZX04014-072)

摘　　要：电阻点焊工艺参数的设置对点焊焊接质量有着非常重要的作用,难以建立精确的数学模型。基于此,提出一种将Logistic映射微粒群优化算法(LPSO)与BP神经网络相结合的方法,对0.8mm厚08AL钢板点焊工艺参数建模优化。在详细分析点焊工艺的基础上,利用BP神经网络建立点焊工艺参数与焊接质量之间的模型,同时结合LPSO的全局寻优能力,对点焊工艺参数进行优化,获得三大主要工艺参数(点焊时间、点焊电流与电极压力)的最优匹配。以点焊时间9周波、点焊电流11.41k N、电极压力1.7k N的最优工艺参数组合进行试验,结果表明,与BP+COA和正交实验法相比,该方法具有更高的可靠性。Resistance spot welding parameters setting plays a very important role in the quality of spot weld- ing, it is difficult to establish a precise mathematical model. Based on this , proposes a method of the Logistic map particle swarm optimization algorithm （LPSO） combined with BP neural network, to model and opti- mize the spot welding process parameters of 0.8mm thick 08AL steel plate. Based on the detailed analysis of spot welding process, using BP neural network to build model between welding parameters and welding quality,optimize the welding parameters combining LPSO global optimization capability, accessed to the three main process parameters （spot welding time, welding current and electrode pressure） best match. Exper- iment the optimal process parameters of spot welding 9cycle, spot welding current 11.41KA, electrode pressure 1.7kN, the result show that compared with the BP ＋ COA and orthogonal experimental method, this method is more reliable.

关 键 词：LPSO BP神经网络 电阻点焊 工艺参数优化 

分 类 号：TH162[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]