基于改进乌鸦算法优化SCN的TE过程故障诊断  

TE process fault diagnosis based on improved crow algorithm optimizing SCN

作　　者：赵文虎 王文 梁晏宾 ZHAO Wenhu;WANG Wen;LIANG Yanbin(Department of Mechanical and Electrical Engineering,Xinjiang Industrial Vocational and Technical College,Urumqi 830022,China)

机构地区：[1]新疆工业职业技术学院机电工程系,乌鲁木齐830022

出　　处：《智能计算机与应用》2024年第6期153-157,共5页Intelligent Computer and Applications

摘　　要：现代化工过程变得越来越复杂,及时准确发现故障显得尤为重要。针对随机配置网络(SCN)泛化能力差、分类准确率低以及乌鸦搜索算法(CSA)寻优能力差等问题,提出了改进乌鸦算法(ICSA)优化随机配置网络的田纳西-伊斯曼(TE)过程故障诊断方法,结合ICSA算法和其它优化算法的寻优对比试验以及优化SCN网络对TE过程中不同故障的分类结果,可以得出,将所提方法应用于TE过程中,可以明显提高不同故障的分类准确率,整个测试集的分类准确率也高达97.6%,具有较好的分类效果,也更符合现代化工生产的需求。Modern chemical processes are becoming more and more complex,so it is particularly important to find faults timely and accurately.Aiming at the problems of poor generalization ability,low classification accuracy of Stochastic Configuration Network and poor optimization ability of Crow Search Algorithm,the Tennessee-Eastman process fault diagnosis method of Improving Crow Search Algorithm to optimize Stochastic Configuration Network is proposed.Combined with the optimization comparison test of Improved Crow Search Algorithm and other optimization algorithms and the classification results of different faults in TE process by optimizing Stochastic Configuration Network,it can be concluded that,applying the proposed method to the TE process can significantly improve the classification accuracy of different faults,and the classification accuracy of the whole test set is also up to 97.6%,which has a good classification effect and is more in line with the requirements of modern chemical production.

关 键 词：化工过程 随机配置网络 改进乌鸦算法 田纳西-伊斯曼过程 分类准确率 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]