摘要
将模糊神经网络用于冲天炉铁液质量预测,构造了一个5层前馈网络,利用27组实验数据对网络进行训练,并对4组未知样本进行预测。结果表明,与目前所用热分析法相比,该网络模型在处理铁液质量这类在一定程度上具有不确定性的多变量非线性对象方面,能够消除建立模型时人为限定,提高预测精度;有效处理模糊信息,而且具有较强的学习能力,适应能力和泛化能力。
In this paper, a fuzzy neural network, with a five-storey structure back propagation network has been developed for predicting the quality of cupola molten iron. 27 groups of the experimental samples were used to train the network and 4 groups are used to verify the network. The results show that for predicting the quality of molten iron with uncertainty and the multi-variable nonlinear object, to some extent, a fuzzy neural network model is better than conventional thermal analysis methods to eliminate the factitious limited on modeling and to improve the prediction accuracy and effective, and to be easy to deal with fuzzy information. It is shown that the method is feasible and has strong learning ability, generalization ability and adaptability.
出处
《铸造技术》
CAS
北大核心
2010年第4期487-490,共4页
Foundry Technology
关键词
冲天炉铁液
模糊神经网络
质量预测
Cupola molten iron
Fuzzy neural network
Predicting the quality
