神经网络应用于烧结矿质量在线推断

On-line Inference of Sintering Quality via Neural Networks

摘要: 针对烧结过程生产实际,运用神经网络中的BP学习算法设计了分类器,用于在线推断烧结矿的质量。为了加快BP学习算法的收敛速度,采用了自适应变步长学习算法。实验结果表明,由此建立的烧结过程神经网络质量预报模型,预报正确率高,具有很好的泛化能力。

Abstract: Presents a new method of on-line inference of sintering quality. Neural networks to build the sinternig quality inference model are used. To speed the learning, a fast BP learning algorithm with adaptive variable step size via linear reinforcement is presented.The experiment result is satisftory, and this method may be used widely in other complicated production proasses.