摘要
应用单输入、输出单元的BP人工神经网络模型算法,对在1450℃实验测得的Fe-V(Nb)-C系中碳的饱和溶解度热力学数据进行非线性拟合分析。其拟合误差检验精度比文献犤1犦线性回归精度更高、更能体现出实验数据的原始变化规律及排除偶然因素的干扰。该方法可供金属熔体热力学性质研究时参考。
The BP manual neural network model of single input and single output was used to make nonlinear imi-tation analysis of carbon solubilit y(in Fe -V-C and Fe -Nb -C systems )measured experimentally at 1450℃.The error examine exactness of this method is better than linear regression method on.This treatment well describes the changing law of experimental data and gets rid of disturban ces caused by accidental factors.This treatment method can be refered to during researching the thermodynamic s of metal melts.
出处
《安徽工业大学学报（自然科学版）》
CAS
2002年第4期284-287,共4页
Journal of Anhui University of Technology（Natural Science）
