【摘要】 良好的焊道是成功进行电弧增材制造的保障,其受到焊接电流、电压、扫描速度、送丝速度等多种参数影响。提出了以焊道高度、宽度为形貌特征的4输入2输出BP神经网络模型,并利用PSO进行了神经网络权值的优化求解。实验结果表明,设计的BP神经网络实现了对焊道形貌的预测,为后续电弧增材制造的实时预测与控制提供了模型基础。更多还原

【Abstract】 A good weld bead provides a guarantee for successful wire arc additive manufacturing,which is affected by various parameters such as welding current,voltage,scanning speed,and wire feed speed. A four-input and two-output BP neural network model is proposed,which is applied to the surface morphology characteristic identification of the weld bead height and width. The particle swarm optimization( PSO) algorithm is used to optimize the neural network weight. Experimental results show that the BP neural network designed in this paper realizes the prediction of the weld bead morphology and that it provides a model basis for the real-time prediction and control of subsequent arc additive manufacturing.更多还原