摘要
利用人工神经网络技术建立了结构钢端淬曲线预测系统的数学模型 ,该模型覆盖了较大范围的钢种 ,系统除了能对已训练过的钢种的端淬曲线准确描述外 ,还能在一定精度范围内对新钢种的端淬曲线进行预测。本文还研究了训练步数。
The mathematical model is established by making use of ANN technology. Many kinds of end\|quench curves of steels are included in this model, and the model can be used to predict the steel Jominy hardness profiles in definite range of chemical composition. Nearly 300 sets of data are collected and the ANN model is trained by using these data. The paper also discuses the influence of training steps and number of training data on predicting precision.
关键词
人工神经网络
数学模型
端淬曲线
结构钢
Artifical Neural Networks(ANN)
\ mathematical model
\ end\|quench curves
\ prediction
