异种钢焊接接头高温时效碳迁移的系统辨识与预测

System identification and predication of carbon migration in dissimilar metal weldments on high-temperature based

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 。

According to properties of both mother metals of austenitic stainless steel and ferritic heat resisting steel and welding procedure, applying the least square method, the finite difference calculation and the method of the BPN's studying algorithm of random iteration of artificial neural network(ANN) to the nonlinear model and state identification of carbon diffusion of dissimilar metal weldments (DMWs) was consulted, analyzed and compared. In order to identificate and predict carbon migration in DMWs on high temperature based, an improved calculating method is put forward.