基于局部线性嵌入的滑动窗口故障检测方法  

Fault detection method for sliding window based on local linear embedding

在线阅读下载全文

作  者:冯立伟[1] 艾浩 孟天祥 李元[3] FENG Liwei;AI Hao;MENG Tianxiang;LI Yuan(College of Science,Shenyang University of Chemical Technology,Shenyang 110142,China;College of Computer Science and Technology,Shenyang University of Chemical Technology,Shenyang 110142,China;College of Information Engineering,Shenyang University of Chemical Technology,Shenyang 110142,China)

机构地区:[1]沈阳化工大学理学院,辽宁沈阳110142 [2]沈阳化工大学计算机科学与技术学院,辽宁沈阳110142 [3]沈阳化工大学信息学院,辽宁沈阳110142

出  处:《传感器与微系统》2025年第3期148-151,共4页Transducer and Microsystem Technologies

基  金:国家自然科学基金资金项目(62273242)。

摘  要:针对连续搅拌反应釜(CSTR)中传感器发生故障难以检测的问题。提出了一种基于局部线性嵌入的滑动窗口(LLE-SW)故障检测方法。首先,LLE-SW提取训练数据的非线性特征并实现维数的约简;其次,通过构造新样本与训练数据的损失函数,并使其最小化,获得新样本的权重矩阵,解决了传统LLE方法没有投影矩阵、新样本无法投影的困难。采用加权平均距离平方和D^(2)作为统计量避免了数据不符合统计量T^(2)要求高斯分布的前提条件,从而发生故障的漏报。最后,通过引入滑动窗口,放大故障的偏移幅值,显著提高了D^(2)的检测效果。CSTR的实验结果表明,所提方法与主成分分析(PCA)、核主成分分析(KPCA)和LLE相比,检测效果更加出色,可作为故障检测领域的有效方法。Aiming at the problem that it is difficult to detect the failure of sensors in continuous stirred tank reactor(CSTR),a sliding window based on locally linear embedding(LLE-SW)fault detection method is proposed.Firstly,LLE-SW extracts the nonlinear features of the training data and achieves dimensionality reduction.Secondly,the weight matrix of the new sample is obtained by constructing the loss function of the new sample and the training data and minimizing it.Solve the difficulty that traditional LLE methods have no projection matrix and new samples cannot be projected.Using the weighted average distance sum of squares D^(2)as a statistic avoids the prerequisite that the data do not meet the Gaussian distribution required by the statistic T^(2),and thus the underreporting of faults.Finally,the detection effect of the D^(2) is significantly improved by introducing a sliding window to amplify the offset amplitude of faults.The experimental results of CSTR show that the detection effect of the proposed method is better than PCA,KPCA and LLE,and can be used as an effective method in the field of fault detection.

关 键 词:故障检测 局部线性嵌入 连续搅拌反应器过程 滑动窗口 

分 类 号:TP277[自动化与计算机技术—检测技术与自动化装置]

 

参考文献:

正在载入数据...

 

二级参考文献:

正在载入数据...

 

耦合文献:

正在载入数据...

 

引证文献:

正在载入数据...

 

二级引证文献:

正在载入数据...

 

同被引文献:

正在载入数据...

 

相关期刊文献:

正在载入数据...

相关的主题
相关的作者对象
相关的机构对象