摘要
为了保障化工系统的安全运行和高质量生产,准确判别化工过程的故障就显得尤为重要。针对田纳西-伊斯曼(Tennessee Eastman,TE)过程故障难以区分以及神经网络等方法在故障诊断中分类准确率较低、分类不稳定等问题,提出一种优化改进极限学习机(extreme learning machine,ELM)的TE过程故障诊断模型。首先利用核主成分分析(kernel principal components analysis,KPCA)方法对TE过程数据进行降维和特征提取,然后采用改进蝴蝶算法(improved butterfly optimization algorithm,IBOA)优化ELM的权值和阈值,最后利用自适应提升(adaptive boosting,AdaBoost)算法对分类器进行集成,完成故障分类。仿真结果表明,IBOA比其他优化算法具有更好的寻优能力,改进效果显著,AdaBoost-IBOA-ELM模型能够对测试集中的不同故障进行准确分类,最后的分类准确率高达98.5%,通过和其他网络对比,进一步验证了模型的合理性和优越性。
In order to ensure the safe operation and high-quality production of chemical systems,it is particularly important to accurately identify the faults of chemical processes.In order to solve the problems of Tennessee Eastman(TE)process fault indistinguishability,neural network and other methods in fault diagnosis,such as low classification accuracy and unstable classification,a TE process fault diagnosis model with optimized and improved extreme learning machine(ELM)was proposed.Firstly,the kernel principal components analysis(KPCA)method was used to reduce the dimensionality and extract features of the TE process data,then the improved butterfly optimization algorithm(IBOA)was used to optimize the weights and thresholds of the ELM,and finally the adaptive boosting algorithm integrates the classifier to complete the fault classification.The simulation results show that IBOA has better optimization ability than other optimization algorithms,and the improvement effect is significant,and the AdaBoost-IBOA-ELM model can accurately classify different faults in the test set,and the final classification accuracy is as high as 98.5%,which further verifies the rationality and superiority of the model by comparing with other networks.
作者
赵文虎
蔡生宏
王文
ZHAO Wenhu;CAI Shenghong;WANG Wen(Department of Electrical and Power Engineering,Xinjiang Industrial Vocational and Technical College,Xinjiang Urumqi 830022,China)
出处
《工业仪表与自动化装置》
2024年第4期102-109,共8页
Industrial Instrumentation & Automation
基金
新疆乌鲁木齐市工业机器人系统操作员技能大师工作室项目。
关键词
田纳西-伊斯曼过程
核主成分分析
改进蝴蝶算法
极限学习机
故障分类
Tennessee Eastman process
kernel principal component analysis
improved butterfly algorithm
adaptive boosting algorithm
fault classification
