基于局部 ? 整体相关特征的多单元化工过程分层监测

针对一类多单元化工过程的监测问题, 提出基于局部?整体相关特征的分层故障检测与故障定位方法, 通过表征单元内部变量相关性、单元与单元间相关性、局部单元与过程整体相关性, 对过程运行状态进行判断, 以提升过程监测的准确性与可靠性. 首先, 采用典型相关分析, 通过引入邻域单元相关变量提取每个单元的独有特征和外部相关特征; 其次, 对每个单元的独有特征和所有单元的外部相关特征建立统计模型实现分层故障检测; 然后, 建立单元?变量分层贡献图, 对故障单元以及故障变量实现分层定位. 通过在Tennessee Eastman仿真过程和一个实验室级甘油精馏过程中的应用说明所提分层监测方法的有效性.     

一种新颖的深度因果图建模及其故障诊断方法

为了实现复杂工业过程故障检测和诊断一体化建模, 提出了一种新颖的深度因果图建模方法. 首先, 利用循环神经网络建立深度因果图模型, 将Group Lasso稀疏惩罚项引入到模型训练中, 自动地检测过程变量间的因果关系. 其次, 利用模型学习到的条件概率预测模型对每个变量建立监测指标, 并融合得到综合指标进行整体工业过程故障检测. 一旦检测到故障, 对故障样本构建变量贡献度指标, 隔离故障相关变量, 并通过深度因果图模型的局部因果有向图诊断故障根源, 辨识故障传播路径. 最后, 通过田纳西?伊斯曼过程进行仿真验证, 实验结果验证了所提方法的有效性.     

Reconstruction-based multivariate contribution analysis for fault isolation: A branch and bound approach

Identification of faulty variables is an important component of multivariate statistical process monitoring (MSPM); it provides crucial information for further analysis of the root cause of the detected fault. The main challenge is the large number of combinations of process variables under consideration, usually resulting in a combinatorial optimization problem. This paper develops a generic reconstruction based multivariate contribution analysis (RBMCA) framework to identify the variables that are the most responsible for the fault. A branch and bound (BAB) algorithm is proposed to efficiently solve the combinatorial optimization problem. The formulation of the RBMCA does not depend on a specific model, which allows it to be applicable to any MSPM model. We demonstrate the application of the RBMCA to a specific model: the mixture of probabilistic principal component analysis (PPCA mixture) model. Finally, we illustrate the effectiveness and computational efficiency of the proposed methodology through a numerical example and the benchmark simulation of the Tennessee Eastman process.     

Efficient faulty variable selection and parsimonious reconstruction modelling for fault isolation

Reconstruction-based fault isolation, which explores the underlying fault characteristics and uses them to isolate the cause of the fault, has attracted special attention. However, it does not explore how the specific process variables change and which ones are most significantly disturbed under the influences of abnormality; thus, it may not be helpful to understanding the specifics of the fault process. In the present work, an efficient faulty variable selection algorithm is proposed that can detect the significant faulty variables that cover the most common fault effects and thus significantly contribute to fault monitoring. They are distinguished from the general variables that are deemed to follow normal rules and thus are uninformative to reveal fault effects. To further reveal the fault characteristics, the selected significant faulty variables are then chosen to obtain a parsimonious reconstruction model for fault isolation in which relative analysis is performed on these selected faulty variables to explore the relative changes from normal to fault condition. The faulty variable selection can not only focus more on the responsible variables but also exclude the influences of uninformative variables and thus probe more effectively into fault effects. It can also help in finding a more interesting and reliable model representation and better identify the underlying fault information. Its feasibility is illustrated with simulated faults using data from the Tennessee Eastman (TE) benchmark process.     

Probabilistic contribution analysis for statistical process monitoring: A missing variable approach

Probabilistic models, including probabilistic principal component analysis (PPCA) and PPCA mixture models, have been successfully applied to statistical process monitoring. This paper reviews these two models and discusses some implementation issues that provide alternative perspective on their application to process monitoring. Then a probabilistic contribution analysis method, based on the concept of missing variable, is proposed to facilitate the diagnosis of the source behind the detected process faults. The contribution analysis technique is demonstrated through its application to both PPCA and PPCA mixture models for the monitoring of two industrial processes. The results suggest that the proposed method in conjunction with PPCA model can reduce the ambiguity with regard to identifying the process variables that contribute to process faults. More importantly it provides a fault identification approach for PPCA mixture model where conventional contribution analysis is not applicable.     

Online monitoring and fault identification of mean shifts in bivariate processes using decision tree learning techniques

With modern data collection system and computers used for on-line process monitoring and fault identification in manufacturing processes, it is common to monitor more than one correlated process variables simultaneously. The main problems in most multivariate control charts (e.g., T ² charts, MCUSUM charts, MEWMA charts) are that they cannot give direct information on which variable or subset of variables caused the out-of-control signals. A Decision Tree (DT) learning based model for bivariate process mean shift monitoring and fault identification is proposed in this paper under the assumption of constant variance-covariance matrix. Two DT classifiers based on the C5.0 algorithm are built, one for process monitoring and the other for fault identification. Simulation results show that the proposed model can not only detect the mean shifts but also give information on the variable or subset of variables that cause the out-of-control signals and its/their deviate directions. Finally a bivariate process example is presented and compared with the results of an existing model.     

基于小波变换核主元分析和多支持向量机的过程监控

针对过程工业数据中所含的噪声和干扰信号、过程工业的非线性及基于主元分析(Principal Component Analysis,PCA)的统计性能监控法由于不用过程机理模型的信息从而对故障诊断问题难以在理论上作系统分析的缺陷,提出基于小波变换核主元分析和多支持向量机的过程监控方法,该方法首先采用基于小波变换的收缩阈值去噪法对建模数据进行预处理,以有效抑制过程数据中所含的噪声和干扰信号,然后利用核主元分析来进行故障特征的提取,从而提高非线性统计过程监控的准确性;最后提出多支持向量机用来对故障的来源进行分类,以避免求解核主元空间到原始空间的逆映射.将该方法应用到对TE(Tennessee Eastman,TE)过程的监控,表明了所提出方法的有效性,为过程的监控和故障诊断提供了一个新的方法.     

Variable reconstruction and sensor fault identification using canonical variate analysis

The detection and identification of faults in dynamic continuous processes has received considerable recent attention from researchers in academia and industry. In this paper, a canonical variate analysis (CVA)-based sensor fault detection and identification method via variable reconstruction is described. Several previous studies have shown that CVA-based monitoring techniques can effectively detect faults in dynamic processes. Here we define two monitoring indices in the state and noise spaces for fault detection and, for sensor fault identification, we propose three variable reconstruction algorithms based on the proposed monitoring indices. The variable reconstruction algorithms are based on the concepts of conditional mean replacement and object function minimization. The proposed approach is applied to a simulated continuous stirred tank reactor and the results are compared to those obtained using the traditional dynamic monitoring technique, dynamic principal component analysis (PCA). The results indicate that the proposed methodology is quite effective for monitoring dynamic processes in terms of sensor fault detection and identification.     

基于核典型相关性-熵成分分析的工业过程质量监测方法

工业过程多变量、数据高维度和非线性的特点使得对其质量监测及质量相关的故障诊断变得复杂.融合核熵成分分析(KECA)及典型相关分析(CCA)方法的思想,进行特征提取降维的同时确保所提取特征与质量变量的最大相关性,提出一种新的质量相关的工业过程故障检测方法.首先,采用KECA对输入数据进行核空间的映射及特征提取,同时融合CCA算法思想使得所提取特征与质量变量间关联最大化;然后,构建监测统计量并用Parzen窗估计其控制限,用于过程的故障检测;最后,运用所提方法对带钢热连轧工业过程实际生产数据进行分析,并与其他4种传统非线性算法对比分析,实验结果验证了所提方法的准确性、有效性及先进性.     

回转窑烧成系统故障诊断方法研究

针对水泥烧成系统过程变量繁多、变量间静态关系耦合强等特点,采用因子分析方法建立静态过程监控模型。针对系统时序相关问题,结合经典动态主元分析DPCA方法和典型变量分析CVA方法,提出典型变量动态主元分析CVDPCA过程监控方法,有效解决了DPCA方法扩展后的数据矩阵维度大等不足之处。将算法用于水泥烧成系统故障检测,结果表明该算法能准确识别故障和更早检测到微小渐变故障。将CVA和DPCA算法相结合,可以同时监控动态过程和静态关系,且不需要大量的故障数据建立故障模型池,具有一定研究价值。     

Progressive multi-block modelling for enhanced fault isolation in batch processes

A multi-block progressive modelling approach is proposed for enhanced fault isolation in batch processes. The unfolding of batch data typically leads to matrices with a large number of columns and this complicates contribution analysis. In order to rapidly focus fault isolation in batch processes, it would be desirable to employ multi-block modelling techniques. Multi-block model such as consensus principal component analysis (CPCA) can produce multiple monitoring charts for sub-blocks and block loadings and block scores can be obtained which can represent unique behaviour of each sub-block. CPCA model uses super score which is the same as score from normal principal component analysis (PCA) model and it does not produce enhanced monitoring performance. Multi-block PCA (MBPCA) model using block score for model calculation can represent sub-blocks’ character but block scores are obtained from super loading so it may not be the best way to describe sub-blocks. A new MBPCA model is proposed for better expression of each sub-block. Through progressive modelling and contribution analysis, variables related to or affected by the fault, as well as the associated time information, are gradually identified. This enables a fault propagation path being established. The proposed method is applied to a benchmark simulated penicillin production process, PenSim.     

基于自联想核回归的带离群值化工过程故障检测EI北大核心CSCD

基于数据驱动的故障检测模型通常要求训练数据必须是正常操作条件下的测量值.然而在实际工业生产过程中,即使在正常工况下,数据集中也难以避免存在离群值.此时若仍采用传统的基于多元统计分析的方法,其监测模型的控制限会受到严重影响,造成故障漏报.因此,为了确保当训练数据包含离群值时,监测模型仍然呈现较好的故障检测效果,本文提出了一种基于自联想核回归的故障检测方法.首先基于最小化β散度的鲁棒预白化算法对训练集进行白化计算,消除变量之间相关性对样本相似度度量的影响.然后通过自联想核回归算法重构正常工况下的验证数据,根据重构误差建立模型监测指标.为了消除离群值对故障样本重构的影响,构造截断函数来避免离群样本参与相似故障数据的重构,并对所有参与构建Q统计量的残差变量基于指数加权滑动平均方法自适应加权,得到新的监测统计量.将该方法运用于田纳西–伊斯曼过程并与其他方法进行比较,验证了本文所提故障检测算法的有效性.     

Fault detection in dynamic systems using the Kullback–Leibler divergence

This paper proposes detecting incipient fault conditions in complex dynamic systems using the Kullback–Leibler or KL divergence. Subspace identification is used to identify dynamic models and the KL divergence examines changes in probability density functions between a reference set and online data. Gaussian distributed process variables produce a simple form of the KL divergence. Non-Gaussian distributed process variables require the use of a density-ratio estimation to compute the KL divergence. Applications to recorded data from a gearbox and two distillation processes confirm the increased sensitivity of the proposed approach to detect incipient faults compared to the dynamic monitoring approach based on principal component analysis and the statistical local approach.     

Improved principal component monitoring using the local approach

This paper shows that current multivariate statistical monitoring technology may not detect incipient changes in the variable covariance structure nor changes in the geometry of the underlying variable decomposition. To overcome these deficiencies, the local approach is incorporated into the multivariate statistical monitoring framework to define two new univariate statistics for fault detection. Fault isolation is achieved by constructing a fault diagnosis chart which reveals changes in the covariance structure resulting from the presence of a fault. A theoretical analysis is presented and the proposed monitoring approach is exemplified using application studies involving recorded data from two complex industrial processes.     

Integrating independent component analysis and support vector machine for multivariate process monitoring

This study aims to develop an intelligent algorithm by integrating the independent component analysis (ICA) and support vector machine (SVM) for monitoring multivariate processes. For developing a successful SVM-based fault detector, the first step is feature extraction. In real industrial processes, process variables are rarely Gaussian distributed. Thus, this study proposes the application of ICA to extract the hidden information of a non-Gaussian process before conducting SVM. The proposed fault detector will be implemented via two simulated processes and a case study of the Tennessee Eastman process. Results demonstrate that the proposed method possesses superior fault detection when compared to conventional monitoring methods, including PCA, ICA, modified ICA, ICA–PCA and PCA–SVM.     

Nonlinear process monitoring using kernel dictionary learning with application to aluminum electrolysis process

In practice, because of complex mechanism processes, such as heating process, volume heterogeneity, and various chemical reaction characteristics, there is a nonlinear relationship among variables in industrial systems. The nonlinearity brings some difficulties to process monitoring. In order to ensure that the process monitoring system can work normally in nonlinear production processes, the nonlinear relationship between variables ought to be considered. In this work, a new fault detection and isolation method based on kernel dictionary learning is presented. In detail, the linearly inseparable data is mapped to a high-dimensional space. Then, a new nonlinear dictionary learning method based on kernel method was proposed to learn the dictionary. After obtaining the dictionary, the control limit can be calculated from the training data according to the kernel density estimation (KDE) method. When new data arrive, they can be represented by the well-learned dictionary, and the kernel reconstruction error can be used as a classifier for process monitoring. As for the fault data, the iterative reconstruction based method is proposed for fault isolation. In order to evaluate the effectiveness of the proposed process monitoring method, some extensive experiments on a numerical simulation, the continuous stirred tank heater (CSTH) process, and a real industrial aluminum electrolysis process are conducted. The proposed method is compared with several state-of-the-art process monitoring methods and the experimental results show that the proposed method can provide satisfactory monitoring results, especially for some small faults, thus it is suitable for process monitoring of nonlinear industrial processes.     

Canonical correlation analysis-based fault detection methods with application to alumina evaporation process

In this paper, canonical correlation analysis (CCA)-based fault detection methods are proposed for both static and dynamic processes. Different from the well-established process monitoring and fault diagnosis systems based on multivariate analysis techniques like principal component analysis and partial least squares, the core of the proposed methods is to build residual signals by means of the CCA technique for the fault detection purpose. The proposed methods are applied to an alumina evaporation process, and the achieved results show that both methods are applicable for fault detection, while the dynamic one delivers better detection performance.     

Generalized grouped contributions for hierarchical fault diagnosis with group Lasso

In process industries, it is necessary to conduct fault diagnosis after abnormality is found, with the aim to identify root cause variables and further provide instructive information for maintenance. Contribution plots along with multivariate statistical process monitoring are standard tools towards this goal, which, however, suffer from the smearing effect and high diagnostic complexity on large-scale processes. In fact, process variables tend to be naturally grouped, and in this work, a novel fault identification strategy based on group Lasso penalty along with a hierarchical fault diagnosis scheme is proposed by leveraging group information among variables. By introducing the group Lasso as a regularization approach, groups of irrelevant variables tend to yield exactly zero contributions collectively, which help find the exact root cause, alleviate the smearing effect, and furnish clear diagnostic information for process practitioners. For online computational convenience, an efficient numerical solution strategy is also presented. Besides, it turns out that the proposed approach also applies to dynamic monitoring models with lagged measurements augmented, thereby enjoying widespread generality. Its effectiveness is evaluated on both the Tennessee Eastman benchmark process and a pilot-scale experiment apparatus.     

基于PCA-RFR的传感器故障定位方法

基于数据驱动的故障诊断方法近些年来得到广泛的研究和应用,但这些方法主要针对于故障检测,对于故障根源的定位尚未得到充分解决。本文提出一种基于主成分分析(PCA)和随机森林回归(PFR)的因果分析故障定位方法(PCA-PFR)。该方法通过将离线故障数据段中的变量作为输入,与之对应的统计量作为输出建立随机森林回归模型,然后通过模型的变量重要性度量来得到过程变量对统计量的因果关系系数,其中值越大的变量被认为越有可能是引起故障发生的故障变量。最后通过一个数值案例和TE过程仿真实验,表明该方法的有效性。     

基于核规范变量分析的非线性故障诊断方法

提出一种基于核规范变量分析（KCVA）的非线性过程故障诊断方法．该方法使用核函数完成非线性空间到高维线性空间的映射,避免了高维空间中的数据处理和非线性映射函数的使用．在线性空间中使用规范变量分析（CVA）来辨识状态空闻模型,从数据中提取状态信息．3个监测量（Tr^2,Ts^2,Q）用来进行故障检测,同时使用贡献图分离故障变量,并判断故障原因．在CSTR系统上的仿真结果表明,KCVA方法比主元分析法（PCA）和CVA方法能更灵敏地检测到故障的发生,更有效地监控过程变化．