Analysis of performance criteria in sampled-data fault detection

Sampled-data control systems are widely used in industry. In this paper the problem of fault detection and isolation (FDI) in sampled-data systems is studied. Many existing methods to design a robust sampled-data FDI are based on optimization of a norm based performance index. Our focus in this study is on the selection of the performance index. It is shown that the existing performance indices are not appropriate choices in the sense that they do not satisfy some expected intuitive properties. To resolve this, an alternative performance index is defined after converting the FDI problem to a standard control problem. This performance index is shown to satisfy the expected properties.     

Norm invariant discretization for sampled-data fault detection

In this paper, the problem of fault detection in sampled-data systems is studied. It is shown that norms of a sampled system are equal to the corresponding norms of a certain discrete time system. Based on this discretization, the sampled-data fault detection problem can be converted to an equivalent discrete-time problem. A framework that unifies the H₂ and H_∞ optimal residual generators in sampled-data systems is then proposed.     

Residual generation and evaluation of networked control systems subject to random packet dropout

In this paper, we address fault detection for networked control systems subject to random packet dropout. The packet dropout is assumed to be existing in the sensor-to-controller link and the controller-to-actuator link. Both parity space and observer based residual generation and evaluation approaches are proposed. In parity space based fault detection scheme, a new optimization index is proposed to deal with stochastic system parameters caused by random packet dropout, while in observer based scheme, this is accomplished by introducing a reference model. In order to evaluate performance of the designated threshold, the corresponding false alarm rate is given. The two fault detection schemes can ensure both robustness to packet dropout as well as disturbance and sensitivity to fault. An experimental study is employed to verify that the proposed method performs better than the existing approaches.     

A frequency domain approach to fault detection in sampled-data systems

This paper deals with fault detection problems in sampled-data (SD) systems. A tool is first introduced for the analysis of intersample behavior of SD systems in the frequency domain from the viewpoint of fault detection and isolation. Based on it, a direct design approach of fault detection systems for SD systems is proposed, and further the problem of full decoupling from unknown disturbances is studied.     

Design of unknown input functional observers for nonlinear systems with application to fault diagnosis

This paper considers the design of low-order unknown input functional observers for robust fault detection and isolation of a class of nonlinear Lipschitz systems subject to unknown inputs. The proposed functional observers can be used to generate residual signals to detect and isolate actuator faults. By using the generalized inverse approach, the effect of the unknown inputs can be decoupled completely from the residual signals. Conditions for the existence and stability of reduced-order unknown input functional observer are derived. A design procedure for the generation of residual signals to detect and isolate actuator faults is presented using the proposed unknown-input observer-based approach. A numerical example is given to illustrate the proposed fault diagnosis scheme in nonlinear systems subject to unknown inputs.     

Semi-supervised kernel canonical correlation analysis with application to human fMRI

Kernel canonical correlation analysis (KCCA) is a general technique for subspace learning that incorporates principal components analysis (PCA) and Fisher linear discriminant analysis (LDA) as special cases. By finding directions that maximize correlation, KCCA learns representations that are more closely tied to the underlying process that generates the data and can ignore high-variance noise directions. However, for data where acquisition in one or more modalities is expensive or otherwise limited, KCCA may suffer from small sample effects. We propose to use semi-supervised Laplacian regularization to utilize data that are present in only one modality. This approach is able to find highly correlated directions that also lie along the data manifold, resulting in a more robust estimate of correlated subspaces.Functional magnetic resonance imaging (fMRI) acquired data are naturally amenable to subspace techniques as data are well aligned. fMRI data of the human brain are a particularly interesting candidate. In this study we implemented various supervised and semi-supervised versions of KCCA on human fMRI data, with regression to single and multi-variate labels (corresponding to video content subjects viewed during the image acquisition). In each variate condition, the semi-supervised variants of KCCA performed better than the supervised variants, including a supervised variant with Laplacian regularization. We additionally analyze the weights learned by the regression in order to infer brain regions that are important to different types of visual processing.     

A novel ensemble construction method for multi-view data using random cross-view correlation between within-class examples

Correlated information between multiple views can provide useful information for building robust classifiers. One way to extract correlated features from different views is using canonical correlation analysis (CCA). However, CCA is an unsupervised method and can not preserve discriminant information in feature extraction. In this paper, we first incorporate discriminant information into CCA by using random cross-view correlations between within-class examples. Because of the random property, we can construct a lot of feature extractors based on CCA and random correlation. So furthermore, we fuse those feature extractors and propose a novel method called random correlation ensemble (RCE) for multi-view ensemble learning. We compare RCE with existing multi-view feature extraction methods including CCA and discriminant CCA (DCCA) which use all cross-view correlations between within-class examples, as well as the trivial ensembles of CCA and DCCA which adopt standard bagging and boosting strategies for ensemble learning. Experimental results on several multi-view data sets validate the effectiveness of the proposed method.     

MVP:基于CCA的多视图数据相关性预测方法

多视图的数据广泛存在于真实的应用中。比如说网络上用户标注的图像,一个视图是由图像的底层特征去表征,而另一个则由文本特征去表征。如何从这种类型的数据中有效地挖掘出有价值的信息对于做数据挖掘和数据检索的人来说具有很大的挑战性。提出多视图的预测算法(MVP)去获取一个子空间,在这个子空间上,通过典型相关分析使得两个视图之间的相互关系最大化。在训练步,期望能通过学习同时得到典型向量组成的子空间和对应典型向量的相关系数;在预测步,通过把数据投影到子空间上产生多视图数据的得分向量。再由得分向量通过多重回归有效地判断出测试样本两个视图之间是不是存在相互关系。基于文本标注图像的实验表明了算法的有效性。     

Design of parametric fault detection systems： An H-infinity optimization approach

Problems related to the design of observer-based parametric fault detection (PFD) systems are studied. The core of our study is to first describe the faults occurring in system actuators, sensors and components in the form of additive parameter deviations,then to transform the PFD problems into a similar additive fault setup, based on which an optimal observer-based optimization fault detection approach is proposed. A constructive solution optimal in the sense of mininfizing a certain peffomaance index is developed. The main results consist of defining parametric fauk detectability, formulating a PFD optimization problem and its solution.A numerical example to demonstrate the effectiveness of the proposed approach is provided.     

案头牡丹农艺指标与基质配方理化指标关系的研究

采用主成分分析的方法对牡丹农艺指标进行了简化,然后采用典型相关分析法对反映案头牡丹生长状况的一组指标与反映基质理化性质状况的另一组指标进行分析。结果表明案头牡丹的农艺指标与基质理化指标相关极显著,两组指标之间的相关主要是由株花蕾数与容重有显著相关引起的。统计分析提示案头牡丹与基质配方理化性质的差别主要是由于容重造成的。案头牡丹的开花朵数主要与电导率的高低有关。     

Improving heat exchanger supervision using neural networks and rule based techniques

The work is aiming to the supervision of heat exchangers fouling monitoring. The fouling known as deposition of undesirable material on the heat transfer surface degrades the performance of heat exchangers. The fouling of heat exchangers in process plants results in a significant cost impact in terms of production losses, energy efficiency, and maintenance costs. To overcome mentioned inconveniences a novel supervision strategy is proposed, reporting innovative techniques and main results of an application tool to diagnose the heat transfer efficiency of a heat exchanger of a pilot plant using neural network based models and parity space approaches associated to a rule based decision making strategy. The developed strategy is fragmented into several modules connected between them following a causal logic flowchart. The first module checks the consistence of the supervision system. The second module monitories the heat exchanger for fouling condition with the ability to diagnose the probable causes of fouling. A third module predicts the remaining operating time under acceptable conditions, associated to a decision making task to schedule the supervision flowchart.     

Automatic radiometric normalization of multitemporal satellite imagery with the iteratively re-weighted MAD transformation

A recently proposed method for automatic radiometric normalization of multi- and hyperspectral imagery based on the invariance property of the Multivariate Alteration Detection (MAD) transformation and orthogonal linear regression is extended by using an iterative re-weighting scheme involving no-change probabilities. The procedure is first investigated with partly artificial data and then applied to multitemporal, multispectral satellite imagery. Substantial improvement over the previous method is obtained for scenes which exhibit a high proportion of change.