针对奥氏体不锈钢和铁素体类耐热钢两类母材金属的特点及相应的焊接工艺 ,分别对最小二乘法、有限差分法和人工神经网络法 (ANN)中的BP算法在异种焊接接头高温时效碳迁移的系统辨识与预测结果进行了分析和比较 ,提出了用改进的算法进行异种钢焊接接头高温时效碳迁移的状态辨识及预测 According to the characteristics of the two types of base metal, such as austenitic stainless steel and ferritic heat-resistant steel, and corresponding welding processes, the least square method, finite difference method and BP algorithm in artificial neural network (ANN) The system identification and prediction results of the high temperature aging carbon bond of the joint were analyzed and compared, and the identification and prediction of the high temperature aging carbon transfer of the dissimilar steel welded joint by the improved algorithm were proposed