论文部分内容阅读
采用三相次级整流点焊机和动态参数检测系统,研究了硬铝合金LY12直流点焊焊接性;在大量工艺试验基础上,运用神经网络方法,构造了一种离散型人工神经网络点焊质量预测与评估模型。研究结果表明,对点焊焊接电流、极间电压和焊点剪切力等预测模型的输入输出参数进行离散化处理,可以实现输入向量空间到输出向量空间的映射,构造的预测模型具有良好的有效性和容错性,适用于铝合金直流点焊质量的预测和评估,为最终实现电阻点焊的智能制造奠定了基础。
Adopting the three-phase secondary rectifying spot welder and the dynamic parameter detection system, the welding performance of the hard aluminum alloy LY12 DC spot welding was studied. Based on a large number of process experiments, a neural network method was used to construct a discrete artificial neural network spot welding Quality Prediction and Evaluation Model. The results show that the input-output parameters of the spot welding current, the inter-electrode voltage and the shear stress of the welding spot can be discretized so that the input vector space to the output vector space can be mapped. The constructed prediction model has good Validity and fault tolerance, which is suitable for the prediction and evaluation of the quality of DC spot welding of aluminum alloy, which lays the foundation for the intelligent manufacture of resistance spot welding finally.