基于局部邻域标准化和贝叶斯推断的多工况过程监测

发布时间：2018-02-15 05:23

本文关键词： 多工况局部邻域标准化独立成分分析—主元分析贝叶斯推断过程监测　出处：《信息与控制》2017年01期 　论文类型：期刊论文

【摘要】：实际工业过程中往往包含不同运行工况,且每种工况数据一般不服从同一种分布.数据的多分布性和分布的不确定性使得传统的故障诊断方法难以获得满意的效果,因此提出一种基于局部邻域和贝叶斯推断的多工况故障诊断方法.首先,通过局部邻域标准化算法对多工况数据进行预处理;再利用ICA-PCA(independent component analysis and principal component analysis)方法分别对该数据集的高斯特性和非高斯特性进行分析处理,获得全局模型;然后结合贝叶斯推断将多个统计量组合成一个监测统计量,实现多工况过程的在线监测;最后通过数值例子和TE过程的仿真研究,验证了提出方法的可行性和有效性.
[Abstract]:The actual industrial process often contains different operating conditions, and the data of each condition is generally not distributed from the same. The multi-distribution and uncertainty of the data make it difficult for the traditional fault diagnosis methods to obtain satisfactory results. Therefore, a multi-condition fault diagnosis method based on local neighborhood and Bayesian inference is proposed. Firstly, the local neighborhood standardization algorithm is used to preprocess the multi-condition data. Then, the ICA-PCA(independent component analysis and principal component analysis method is used to analyze and process the Gao Si and non-#china_person1# characteristics of the data set, and the global model is obtained, and then the multiple statistics are combined into one monitoring statistic by Bayesian inference. Finally, the feasibility and effectiveness of the proposed method are verified by numerical examples and simulation studies of te process.
【作者单位】：江南大学物联网工程学院自动化研究所;轻工过程先进控制教育部重点实验室江南大学;
【基金】：国家自然科学基金资助项目(21206053,21276111) 江苏省“六大人才高峰”计划资助项目(2013-DZXX-043) 江苏高校优势学科建设工程资助项目(PAPD)
【分类号】：TQ06;TP277

【相似文献】