太原理工大学学报2018，Vol.49Issue(1)：86-93,8.DOI:10.16355/j.cnki.issn1007-9432tyut.2018.01.014
基于高斯模型的工业过程数据的故障预测
Fault Prediction for Industrial Process Based on Gaussian Process Model
摘要
Abstract
A new Gaussian process regression based predictive modeling is proposed to realize fault prediction for the time series data of measured variables in industrial processes.Compared with existing techniques,Gaussian process regression model can show the uncertainty of a result,as well as the confidence interval.By constructing the kernel function of Gaussian process model based on a specific data set,and adding the priori information of industrial process,the performance of the proposed model is further improved.A series of contrastive experiments are conducted on Tennessee Eastman process simulator,better performance is achieved for the proposed fault prediction method.关键词
故障预测/高斯过程回归/田纳西-伊斯曼过程/工业过程数据建模Key words
fault prediction/gaussian process regression/Tennessee Eastman process/industrial process data modeling分类
信息技术与安全科学引用本文复制引用
杨为惠,陈彦萍,温福喜,高聪..基于高斯模型的工业过程数据的故障预测[J].太原理工大学学报,2018,49(1):86-93,8.基金项目
陕西省科技厅工业科技攻关项目(2016GY-092) （2016GY-092）
陕西省科技统筹创新工程-重点产业创新链工业领域项目(2016KTZDGY04-01) （2016KTZDGY04-01）
陕西省教育厅专项科学研究项目(17JK0711,16JK1701) （17JK0711,16JK1701）
工业和信息化部通信软科学研究项目(2017 R-21)资助 （2017 R-21）
