基于人工神经网络的低碳贝氏体钢回火组织预测

Prediction of Tempering Microstructure of Low-carbon Bainite Steel Based on BP Artificial Neural Network

低碳贝氏体钢经控轧控冷后,在不同温度下进行回火处理,获得了不同冷却速度、终冷温度和回火温度下的贝氏体室温组织。结合实验数据和神经网络知识,建立了具有BP算法的人工神经网络,训练结束后的神经网络即成为低碳贝氏体钢回火组织预测模型。误差分析表明,该神经网络模型具有较高的精度,可用于指导低碳贝氏体钢热加工工艺的制定。

The microstructure of low-carbon bainite steel at room temperature was obtained through controlled rolling and cooling and tempering treatment.After analyzing the experimental data of the bainite grain size under different conditions of cooling speed,final cooling temperature and tempering temperature,an artificial neural network with back propagation（BP） algorithm was established and models for predicting the tempering microstructure of low-carbon bainite steel were developed after training.Error analysis shows that the artificial neural network model for bainite microstructure has higher precision,and it can be used for guiding the hot working process of low-carbon bainite steel.