基于阈值故障子空间提取算法的多重故障重构

在统一框架下考虑了多重传感器故障的故障检测与故障重构问题.传感器的多种失效模式给过程监控和故障重构带来了一定的挑战.首先,提出一种可以同时表示加性和乘性传感器故障的故障模型.从数据驱动的角度出发,多重故障可检测性的必要条件和充分条件通过RayleighRitz引理被推导出来.进一步,提出了一种基于阈值故障子空间提取算法的多重故障重构方法.数值仿真表明,所提出的故障重构方法在重构误差指标方面明显优于传统的故障重构方法,验证了文中给出的主要结果.

一种用阈值提取子空间的多步匹配场反演方法

在分析已有的匹配场反演方法的基础上,构造了一种用阈值提取子空间的多步匹配场反演方法。它根据一定反演环境下参数的不同敏感性将参数划分为子集(子空间),并依次在各敏感子空间内反演。反演时用一定的阈值将目标函数优于阈值的参数区域提取出,最后在提取出的已相对缩减的区域和最后一个子空间(通常是不敏感参数子空间)内联合反演全部参数,求得最优值。这样既可减少反演参数空间又能可靠地保证精确度,避免了已有的子空间方法反演结果受非反演参数失配影响的问题。仿真研究结果表明,本算法比已有的两类算法性能上有明显提高。

可行方向算法与模拟退火结合的NMF特征提取方法

NMF子空间特征提取被表示成一个大规模线性约束非线性优化问题 .为了获得更优性能的基图像 ,设计了一个可行方向算法结合模拟退火算法的混合算法来求解这个优化问题 .以基于梯度的可行方向算法作为局部寻优的手段 ,加快收敛速度 ;以模拟退火算法作为全局寻优的手段 ,避免优化过程陷入局部极小点 .同时 ,在模拟退火操作中 ,采用对比度增强算法 ,使获得的基图像更加地空间局部化 .实验表明 ,本文的可行方向算法比采用归一化实现等式约束的原算法在学习的最后阶段有更好的收敛速度 ,所获得的基图像更加地空间局部化 。

DISCRIMINANT INDEPENDENT COMPONENT ANALYSIS AS A SUBSPACE REPRESENTATION

Subspace modeling plays an important role in face recognition. Independent Component Analysis （ICA）, a multivariable statistical analysis technique, can be seen as an extension of traditional Principal Com- ponent Analysis （PCA） technique, which addresses high order statistics as well as second order statistics. In this paper, a new scheme of subspace-based representation called Discriminant Independent Component Analysis （DICA） is proposed, which combines the strength＂ of unsupervised learning of ICA and supcrvised learning of Linear Discriminant Analysis （LDA）, and efficiently enhances the generalization ability of ICA-based representation method. Based on DICA subspace analysis, a set of optimal vectors called ＂discriminant independent faces＂ are learned from face samples. The effectiveness of our method is demonstrated by performance comparisons with some popular methods such as ICA, PCA, and PCA＋LDA. On the large scale database of IIS, significant improvements are observed when there are fewer training samples per person available.

RESEARCH ON TWO-DIMENSIONAL LDA FOR FACE RECOGNITION

The letter presents an improved two-dimensional linear discriminant analysis method for feature extraction. Compared with the current two-dimensional methods for feature extraction, the improved two-dimensional linear discriminant analysis method makes full use of not only the row and the column direc-tion information of face images but also the discriminant information among different classes. The method is evaluated using the Nanjing University of Science and Technology (NUST) 603 face database and the Aleix Martinez and Robert Benavente (AR) face database. Experimental results show that the method in the letter is feasible and effective.

New learning subspace method for image feature extraction

A new method of Windows Minimum/Maximum Module Learning Subspace Algorithm(WMMLSA) for image feature extraction is presented.The WMMLSM is insensitive to the order of the training samples and can regulate effectively the radical vectors of an image feature subspace through selecting the study samples for subspace iterative learning algorithm,so it can improve the robustness and generalization capacity of a pattern subspace and enhance the recognition rate of a classifier.At the same time,a pattern subspace is built by the PCA method.The classifier based on WMMLSM is successfully applied to recognize the pressed characters on the gray-scale images.The results indicate that the correct recognition rate on WMMLSM is higher than that on Average Learning Subspace Method,and that the training speed and the classification speed are both improved.The new method is more applicable and efficient.

A hybrid specific index-related process monitoring strategy based on a novel two-step information extraction method

A two-step information extraction method is presented to capture the specific index-related information more accurately.In the first step,the overall process variables are separated into two sets based on Pearson correlation coefficient.One is process variables strongly related to the specific index and the other is process variables weakly related to the specific index.Through performing principal component analysis(PCA)on the two sets,the directions of latent variables have changed.In other words,the correlation between latent variables in the set with strong correlation and the specific index may become weaker.Meanwhile,the correlation between latent variables in the set with weak correlation and the specific index may be enhanced.In the second step,the two sets are further divided into a subset strongly related to the specific index and a subset weakly related to the specific index from the perspective of latent variables using Pearson correlation coefficient,respectively.Two subsets strongly related to the specific index form a new subspace related to the specific index.Then,a hybrid monitoring strategy based on predicted specific index using partial least squares(PLS)and T2statistics-based method is proposed for specific index-related process monitoring using comprehensive information.Predicted specific index reflects real-time information for the specific index.T2statistics are used to monitor specific index-related information.Finally,the proposed method is applied to Tennessee Eastman(TE).The results indicate the effectiveness of the proposed method.

Dynamic mode decomposition of hairpin vortices generated by a hemisphere protuberance

We present dynamic mode decomposition (DMD) for studying the hairpin vortices generated by hemisphere protuberance measured by two-dimensional (2D) time-resolved (TR) particle image velocimetry (PIV) in a water channel. The hairpins dynamic information is extracted by identifying their dominant frequencies and associated spatial structures. For this quasi-periodic data system, the resulting main Dynamic modes illustrate the different spatial structures associated with the wake vortex region and the near-wall region. By comparisons with proper orthogonal decomposition (POD), it can be concluded that the dynamic mode concentrates on a certain frequency component more effectively than the mode determined by POD. During the analysis, DMD has proven itself a robust and reliable algorithm to extract spatial-temporal coherent structures.