Multimode process monitoring strategy based on improved just-in-time-learning associated with locality preserving projections

In this paper, a multimode process monitoring strategy based on improved just-in-time-learning associated with locality preserving projections (IJITL-LPP) is proposed. First, raw data are projected into the feature space using locality preserving projections (LPP). Second, IJITL searches for similar samples of the query sample in the feature space by introducing a variational inference Gaussian mixture model (VIGMM). Finally, the new statistic named average distance is created to complete process monitoring. In the IJITL, the introduced VI can automatically determine the number of modes, thereby accelerating the efficiency of selecting similar samples. In the process monitoring phase, the average distance can reduce the impact of different mode dispersion on fault detection. In addition, LPP can render the model less sensitive to outliers. Compared with principal component analysis (PCA), LPP, K nearest neighbour rules, Gaussian mixture model (GMM), K-means based-PCA, and just-in-time-learning (JITL)-based LPP, the proposed method has better performance in a numerical case, the Tennessee Eastman process, and the semiconductor etching process.     

斜盘式锥齿少齿差减速器传动件加工新方法

对斜盘式锥齿少齿差行星减速器的一些重要传动元件的制造方案进行了系统研究和论证．主要包括内啮合锥齿轮副、输入轴-斜盘、歪轴的制造及装配、检测工艺等．通过引入内锥齿轮副鼓形齿修正等新工艺理论，为斜盘式锥齿少齿差行星传动的工业应用奠定了基础．     

基于某铀矿的地浸采铀方法

简述了地浸采铀的主要影响因素,提出了一种基于高矿化度砂岩铀矿的地浸采铀工艺——淡化少试剂工艺。此工艺不仅可以有效防止管道堵塞,而且可以提高浸铀的经济效益。     

基于局部保持投影–加权k近邻规则的 多模态间歇过程故障检测策略

针对多模态间歇过程故障检测问题,本文提出一种基于局部保持投影–加权k近邻规则(LPP--Wk NN)的故障检测策略.首先,应用局部保持投影(LPP)方法将原始数据投影到低维主元子空间;接下来,在主元子空间中,应用样本第k近邻的局部近邻集确定每个样本的权重并计算权重统计量Dw;最后,应用核密度估计方法确定Dw控制限并进行故障检测.本文方法应用LPP对过程数据进行维数约减,既能够降低训练过程中离群点对模型的影响,又能够降低在线故障检测的计算复杂度.同时,加权k近邻规则(Wk NN)方法通过引入权重规则能够使得过程故障检测统计量分布具有单模态结构.相比传统的k NN统计量,本文引入的权重统计量具有更高的故障检测性能.通过数值例子和半导体蚀刻过程的仿真实验,并与主元分析(PCA), k NN, Wk NN, LPP--k NN等方法进行比较,实验结果验证了本文方法的有效性.     

A generic approach to recognising isolated, nested and interacting features from 2D drawings

Many useful three-dimensional (3D)-feature recognition systems have been developed over the past few decades. However there is a shortage of feature recognition systems in the two-dimensional (2D) domain. This has an adverse effect on the realisation of CADCAM benefits in Small and Medium Enterprises. Here is an attempt to redress this deficit. This paper presents a hint-based feature recognition system which recognises machining features from 2D orthographic projections through a two-stage process of profile searching and feature completion. Three types of profiles: Type I, II or III are defined in this system. Together they are sufficient to describe most profiles, if not all, in any 2D drawing. Profile searching identifies these profiles while feature completion establishes the cavity emanating from these profiles using a multi-level top–down approach. This process enables the system to deal with the isolated, nested and interacting features using a common framework. The nested features are handled by applying the same reasoning used for isolated features recursively while interacting features are decomposed automatically by the process of profile searching into their constituent isolated features.     

光CT的扇束投影重排方法及投影几何分析

在工业过程监控中,由于被测物理过程的发生时间短,因此光CT装置将不可能象医学CT那样有充足的时间进行"静态"成像,而是必须进行实时性较高的"动态"成像(或称为流动成像).其制约是光线扫描的投影视角和投影数量都不可能太多.根据这一情况,在医学CT中常用的一些图像重建算法如反投影法、滤波反投影法和投影重排方法都必须加以改进才能引用.为此,本文研究扇束投影重排方法在光学流动成像中的应用可行性,分析其计算方法和步骤,并确定光线扇束结构中的光源个数、扇形张角和探测器个数之间的约束关系,同时也分析了内插精度对图像质量的影响.在数值模拟实验中,将投影重排方法与反投影法的重建图象进行了比较,进一步评判本文方法的图象质量和成像实时性等方面的性能.     

On the Computation of Observers in Discrete-Event Systems

The concept of observer was introduced in previous work by the authors on a hierarchical control theory of discrete-event systems (DES). It was shown that the observer property of the “causal reporter” map, which in this theory models information flow in a hierarchical DES, plays a role in ensuring that a nonblocking supervisor at a given level of a hierarchy does not cause blocking in the level below. In this paper, we investigate the following problem: Given a causal reporter map that is not an observer, how can we design an observer by modifying this map? In case the latter is represented by a finite Mealy automaton, an effective computational algorithm is developed for computing an observer with the coarsest possible equivalence kernel that is finer than that of the given map. Three examples are provided for illustration.     

Few views image reconstruction using alternating direction method via ‐norm minimization

In the medical computer tomography field, total variation (TV), which is the ‐norm of the gradient‐magnitude images, is widely used as the regularization based on the compressive sensing theory. To overcome the TV model's disadvantageous tendency of uniformly penalize the image gradient and over smooth the low‐contrast structures, an iterative algorithm based on the ‐norm optimization of the finite difference is proposed. To rise to the challenges introduced by the ‐norm minimization, the algorithm uses the alternating direction method to solve the unconstrained augmented Lagrangian function, which involves a hard thresholding method, a linearization and proximal points technique for each subproblem. The simulation demonstrates the conclusions and indicates that the algorithm proposed in this article can obviously improve the reconstruction quality. © 2014 Wiley Periodicals, Inc. Int J Imaging Syst Technol, 24, 215–223, 2014     

一种改善超分辨率图像重建中边缘质量的方法

超分辨率图像重建技术指通过融合多幅变形、模糊、有噪、频谱混叠的低分辨率降质图像来重建一幅高质量高分辨率图像. 凸集投影 (POCS) 算法是一种广泛使用的超分辨率图像重建方法. 本文提出了一种适用于 POCS 算法的改善高分辨率重建图像边缘质量的方法. 该方法将中心在边缘像素的点扩散函数 (PSF) 与一个指数型权值函数相乘, 使得修改的 PSF 系数沿着边缘正交的方向减小. 实验结果表明, 这样的修改有效地保持了边缘的特性, 明显地提高了重建图像的质量     

500kV变电站实行无人值守运行管理模式的探讨

探讨了国内500kV变电站实行无人值守或少人值守运行模式的可行性。介绍了变电站运行模式改革的方向,比较了国内500kV变电站与220kV变电站、国内外500kV变电站的异同点。以苏州地区500kV变电站为例,从变电站设备缺陷、跳闸事故处理和工作量等几个方面着手,分析了决定变电站运行模式改革的各要素。分析表明,对现有变电站,条件成熟的地区可以实现变电站的少人值守,二次设备能否实现远方操作对变电站实现无人值守至关重要。