摘要
将概率神经网络建模方法与预测思想相结合用于热轧轧制节奏评价,研究并建立了基于PNN神经网络的热轧轧制节奏评价模型。基于涟钢2 250mm热轧厂的实测数据,将建立的PNN网络轧制节奏评价模型用于生产实际,并将结果与BP神经网络进行对比。结果表明,该模型具有便捷、快速、预测精度高、泛化能力强的特点,可代替现有的基于经验公式和经验数据的评价方法,同时为轧制节奏的优化和生产效率的提高提供了参考,具有重要的现实意义。
The probabilistic neural network （PNN） and prediction thought are combined to evaluate mill pacing of hot rolling mill .A model based on PNN is established .To verify the validity of the model and predicted results ,the model is applied to practical production of a hot strip mill and compared with the BP network .The results prove that the model has the advantage of simple structure ,fast calculation ,high prediction accuracy and strong generalization ability ,which is able to substitute for the existing evaluation model based on the empirical formula and empirical da-ta .Meanwhile ,it provides a reference for the optimization of the mill pacing and production efficiency ,so it has an important practical significance .
出处
《中国冶金》
CAS
2014年第3期27-30,共4页
China Metallurgy
关键词
概率神经网络
轧制节奏
评价模型
代替
现实意义
probabilistic neural network
mill pacing
evaluation model
substitution
practical significance
