引用本文: | 曹晨鑫,王昕,王振雷.基于多数据集动态潜变量的在线性能分级评估方法[J].控制理论与应用,2020,37(3):658~666.[点击复制] |
CAO Chen-xin,WANG Xin,WANG Zhen-lei.Online performance grading assessment method based on multiset dynamic latent variables[J].Control Theory and Technology,2020,37(3):658~666.[点击复制] |
|
基于多数据集动态潜变量的在线性能分级评估方法 |
Online performance grading assessment method based on multiset dynamic latent variables |
摘要点击 2170 全文点击 1438 投稿时间:2018-10-26 修订日期:2019-05-31 |
查看全文 查看/发表评论 下载PDF阅读器 |
DOI编号 10.7641/CTA.2019.80833 |
2020,37(3):658-666 |
中文关键词 在线性能分级评估 多数据集动态潜变量 神经网络 动态自相关 乙烯裂解 |
英文关键词 online performance grading assessment multiset dynamic latent variables neural network dynamic auto-correlation ethylene cracking |
基金项目 国家自然科学基金项目(61673268),国家自然科学基金重点项目(61533003), 国家自然科学基金重大项目 (61590922),中央高校基本科研业务费资助(222201814043) |
|
中文摘要 |
针对动态多变量过程中难以提取明确的过程变量的动态关系问题, 本文提出基于多数据集动态潜变量分析(MSDLV)的在线性能分级评估的方法。 首先将过程性能相近的历史数据段划分为不同性能等级的集合, 然后运用 MSDLV 方法提取性能级之间的公共基向量, 保留训练数据中性能相关的过程变化, 将性能相关的特有变化分解为动态部分与静态部分, 提取动态自相关过程的动态因素。 建立动态潜变量与性能等级之间的离线模型, 有效提高在线评估当前过程性能以及判断其所处状态的准确度。 将该方法运用于乙烯裂解炉反应过程, 结果表明该方法具有良好的效果。 |
英文摘要 |
Aiming at the dynamic relationship between process variables in the dynamic multivariate process are implicit and hard to interpret, this paper presents an online performance grading assessment method based on multiset dynamic latent variables(MSDLV). First, similar historical data is sorted into sets with different performance grades. Then common basis vector is obtained by MSDLV algorithm. Performance related variation is extracted from the original data and is divided into dynamic part and static part. The dynamic factors in auto-correlated process are extracted. The offline model is established between latent variables and performance grades. Current performance can be judged online more effectively. Meanwhile, whether process is in the stability grade state or the conversion state can be analysed more accurately. The application result of ethylene cracking process show the method has a good effect. |