基于RBF神经网络的ZGMn13堆焊焊条的设计

The Design of Hardfacing Electrode about ZGMn13 Based on Bayesian Neural Networks

提出了一种基于RBF神经网络的CaO-CaF2-SiO2渣系ZGMn13堆焊焊条配方优化设计方法.利用实验采集的数据对网络进行训练,以加工硬化后的硬度为优化目标,得到最优的焊条配方.实验结果表明:优化后熔敷金属的动载加工硬化性能和静载加工硬化性能良好.

To design a kind of hardfacing electrode about ZGMn13, a method of optimization design about electrode formulation with CaO-CaF2-SiO2 slag system is proposed based on RBF neural networks. The networks were trained with the data collected from experiments. Optimal formulation of electrode was achieved when the hardness after work hardening was considered as the optimal target. The results of the experiment show good performances of both dynamic load hardening and static load hardening.