基于改进模糊ART神经网络的连铸漏钢预报模型  

Breakout Prediction Model Based on Improved Fuzzy ART Neural Network for Continuous Casting Process

作　　者：赵琦[1] 朱苗勇[1] 

机构地区：[1]东北大学材料与冶金学院,辽宁沈阳110004

出　　处：《中国冶金》2007年第10期26-29,53,共5页China Metallurgy

摘　　要：在模糊ART神经网络的基础上,有机结合模糊模式识别和模糊聚类算法,并通过引入新的学习机制和优化网络结构,建立了改进的新型模糊ART神经网络模型;同时,结合某钢厂连铸现场采集的历史数据,将该模型应用于连铸漏钢预报过程中。其结果表明,该模型对粘结漏钢过程中2种典型温度模式的预报率分别达到95.6%和97.8%,报出率都达到100%,且在避免漏报的同时保证了较低的误报率,能准确识别典型的温度模式和预测拉漏事故的发生。Based on fuzzy ART neural network, fuzzy pattern recognition and fuzzy clustering algorithm, an improved fuzzy ART neural network model was presented by introducing a new learning method and optimizing the structure of network. The model was applied to the breakout prediction of continuous casting with history data acquired in a steel work. The results show that the model is effective in identifying two typical temperature patterns of sticking breakout and detecting possible leakages of liquid steel with the prediction rate of 95.6% and 97.8%, respectively. The quote rate of both pattern reach to 100%.

关 键 词：连铸 漏钢预报 模糊神经网络 模糊聚类 

分 类 号：TF341.6[冶金工程—冶金机械及自动化]