Development of a new cluster index for wafer defects

Defect number and defect clustering are two key determinants of wafer yield. Preventing and detecting wafer defects thus is an important issue in integrated circuit manufacturing. Defect clustering tends to grow with increasing wafer size. Methods have been developed for assessing defect clustering on wafers. However, these methods require either statistical assumptions regarding defect distribution or complex computations. This study develops a new cluster index, utilizing the rotating axis technique from multivariate analysis to accurately quantify defect clusters on a wafer. The developed defect-clustering index does not require making assumptions regarding defect distribution. Thus, the proposed method can be efficiently used by engineers with little statistical background. A simulation experiment is conducted to demonstrate the effectiveness of the proposed defect-clustering index.     

Clustering of sources of acoustic-emission signals by the leading-edge rise rate

It is shown that the method of clustering by the leading-edge rise rate of the envelope of an acoustic-emission (AE) signal makes it possible to reduce the time of calculation of the coordinates of flaws as compared to the clustering by shape. The developed technique realizes the on-line localization of massive flows of AE signals with an accuracy sufficient for practical application. The practical parameters of using the method for operation on a steel sheet with a simulator obtained for a statically loaded duralumin specimen and for welding of steel specimens with aluminum and titanium inserts are considered.     

运用改进模糊C均值算法的复杂产品功构单元模块划分方法

为了解决目前已有聚类方法在复杂产品功构单元模块划分过程中存在的问题,提出了一种基于改进模糊C均值算法(D-FCM)的产品功构单元模块划分方法。该方法运用距离矩阵生成样本分层聚类树,结合F统计量方法确定模糊C均值算法最佳聚类数,在此基础上,应用模糊C均值算法进行聚类分析,获得聚类结果。最后,结合实际项目给出该聚类方法在机床模块划分过程中的典型应用,对该方法进行实例验证。以此为基础,开发出机床模块划分系统平台。系统实现及设计结果表明了所提出方法的有效性,为面向配置设计的机床模块划分提供了另一种有效的模块划分方法。     

Optimization of the number of clusters: a case study on multivariate quality control results of segment installation

When encountering too many records, each of which has several attributes, clustering of the data is an important issue on mining and classification. Recently many advances on clustering algorithms have been made such that clustering of data is done precisely and quickly. Clustering algorithms use optimization algorithms which simultaneously provide the number of clusters as default. These algorithms cluster the data so that those which belong to a cluster have maximum similarity and those in different clusters have minimum similarity. The k-means algorithm is a traditional algorithm for clustering problems. One of the most important difficulties of clustering algorithms is determining the number of clusters before starting the algorithm. In other words, by having knowledge on distribution of data, the number of clusters should be estimated and then imported to the problem as an input. In this paper, the data collected on quality control of mechanized tunneling are analyzed. They consist of measurements of 16 characteristics for 200 initial installed rings of segments on the tunnel walls inspected by the quality control team. A dynamic validity index is used and combined to the k-means algorithm for clustering the data so that the optimal number of clusters can be determined simultaneously. The application of the algorithm shows that the total installed rings can be clustered into four clusters. These four classes of quality can best describe the total installed rings on the tunnel in comparison of other number of classes (or clusters). Furthermore, this approach helps the quality team to determine the most effective or best performance executive team whom their installed rings have best class and minimum variations.     

Hierarchical clustering approach for determination of isomorphism among planar kinematic chains and their derived mechanisms

The problem of isomorphism among kinematic chains and their derived mechanism has been a hot area of research from last several years. The researchers so far have proposed many methods which are mainly based on characteristic polynomial and some code based methods to test the isomorphism among kinematic chains. In this present communication a hierarchical clustering based computerized method is proposed for the above said problem and it is tested for planar kinematic chains upto twelve links without any counter examples. In this method a hierarchical clustering algorithm is also developed for the identification of distinct mechanism derived from kinematic chains. In this method kinematic chains are represented in the form of weighted squared shortest path distance matrix and this matrix is further transformed in the form of tree or dendrogram with the help of hierarchical clustering algorithm. This algorithm directly gives the number of distinct mechanism derived from a given kinematic chain. The cophenetic correlation coefficient of dendrogram is used as an index for isomorphism identification among kinematic chains. The proposed method is efficient and accurate and only one matrix for a given kinematic chain is developed for the determination of distinct mechanisms. This method is successfully examined for one degree of freedom, 6, 8, 10, 12 links planar kinematic chains, 9 links two degree of freedom and 10 links three degree of freedom planar kinematic chains. The computer algorithm for the proposed method has been proposed which can easily be converted into a computer program. These results are useful for designers to detect isomorphism in mechanisms derived from kinematic chains and duplication among kinematic chains.     

Contingency table techniques for three dimensional atom probe tomography

A contingency table analysis procedure is developed and applied to three dimensional atom probe data sets for the investigation of fine-scale solute co-/anti-segregation effects in multicomponent alloys. Potential sources of error and inaccuracy are identified and eliminated from the technique. The conventional P value testing techniques associated with chi(2) are shown to be unsatisfactory and can become ambiguous in cases of large block numbers or high solute concentrations. The coefficient of contingency is demonstrated to be an acceptable and useful basis of comparison for contingency table analyses of differently-conditioned materials. However, care must be taken in choice of block size and to maintain a consistent overall composition between experiments. The coefficient is dependent upon block size and solute composition, and cannot be used to compare analyses with significantly different solute compositions or to assess the extent of clustering without reference to that of the randomly ordered case. It is shown that as clustering evolves into larger precipitates and phases, contingency table analysis becomes inappropriate. Random labeling techniques are introduced to infer further meaning from the coefficient of contingency. We propose the comparison of experimental result, mu(exp), to the randomized value, micro(rand), as a new method by which to interpret the quantity of solute clustering present in a material. It is demonstrated that how this method may be utilized to identify an appropriate size of contingency table analysis blocks into which the data set is partitioned to optimize the significance of the results.     

选择高速钢刀具材料的专家模糊聚类算法

在加工中心数控系统内开发的选刀专家系统中应用模糊聚类分析法，结合选刀专家系统，提出了一套新的用于选择高速钢刀具材料牌号的模糊聚类分析模型和算法。实践证明，这种方法适用面广，易于扩充，简单、有效，具有一定的理论意义和较高的实用价值，为选刀专家系统的进一步开发打下了坚实的基础。     

基于聚类分析的航空发动机轴承腔油气两相流型辨识

轴承腔是航空发动机内的重要部件,对其内两相流流动状态的理解有助于提高发动机的润滑效率、提升发动机的工作性能和可靠性。基于流型辨识研究两相流在不同流型下的流动性质是获取轴承腔内物理现象的一个可行途径。在考虑试验获取难易性和辨识能力条件下,结合前人研究成果,应用K-Means聚类法对利于辨识轴承腔3种典型两相流流型的特征参数-无量纲两相介质速度和无量纲压力在近壁面附近的数值进行聚类分析。研究结果表明:该方法能够成功地将轴承腔3种典型流型进行归类,这对基于流型辨识研究轴承腔内3种典型流型下两相介质的物理状态,以及轴承腔润滑设计和换热分析提供了基础。     

Robust clustering approach for the shortest path problem on finite capacity fuzzy queuing network

This paper introduces a new model in the field of fuzzy queuing networks based on the clustering and finite capacity concepts. The proposed model includes fuzzy queuing systems which are located at the nodes of the network. Arc lengths, interarrival times, and service times are all fuzzy triangular fuzzy numbers. In order to find the shortest path on this network, queuing systems should be transformed to waiting times. The waiting times of each system are calculated by a conditional transformation. On the other hand, a robust fuzzy method is proposed for clustering of the arriving customers to the network. Robustness of this method prevents noisy data to affect results. Outputs of the clustering reduce shortest path calculations drastically. Based on a simulation process, the fuzzy queuing network is reduced to a deterministic network. A robust simulated annealing is designed for this network to find the shortest path. Numerical results showed that the clustering process is successful in eliminating outliers, and could be addressed as an efficient method. The convergence and solution time of the algorithm is reasonably better in comparison with published methods. Several experiments are conducted to compare the proposed method with corresponding researches.     

On the Appearance of Lubricant Film Rupture in Cylindrical Journal Bearings

Based on energy and mass balances in conjunction with the classical Reynolds equation in the film region, expressions for cavitation phenomenon are developed. They determine the circumferential location angle where cavitation might start, and the potential number of bifurcating fluid streamer surfaces (boundaries between gas-liquid flow) can be estimated. The expressions depend on the journal angular velocity, equilibrium eccentricity ratio, and bearing and fluid characteristics; they strongly influence cavitation and have not been considered in previous studies. Conditions under which the cavitation phenomenon does not occur are also given. Finally, a comparison to previous research results showing a very close agreement is presented.     

Online automatic identification of the modal parameters of a long span arch bridge

The “Infante D. Henrique” bridge is a concrete arch bridge, with a span of 280 m that crosses the Douro River, linking the cities of Porto and Gaia located in the North of Portugal. This structure is being monitored by a recently installed dynamic monitoring system that comprises 12 acceleration channels. This paper describes the bridge structure, its dynamic parameters identified with a previously developed ambient vibration test, the installed monitoring equipment and the software that continuously processes the data received from the bridge through an Internet connection. Special emphasis is given to the algorithms that have been developed and implemented to perform the online automatic identification of the structure modal parameters from its measured responses during normal operation. The proposed methodology uses the covariance driven stochastic subspace identification method (SSI-COV), which is then complemented by a new algorithm developed for the automatic analysis of stabilization diagrams. This new tool, based on a hierarchical clustering algorithm, proved to be very efficient on the identification of the bridge first 12 modes. The results achieved during 2 months of observation, which involved the analysis of more than 2500 datasets, are presented in detail. It is demonstrated that with the combination of high-quality equipment and powerful identification algorithms, it is possible to estimate, in an automatic manner, accurate modal parameters for several modes. These can then be used as inputs for damage detection algorithms.     

KIFC1 overexpression promotes prostate cancer cell survival and proliferation <i>in vitro</i> by clustering of amplified centrosomes via interaction with Centrin 2

Mitotic kinesin KIFC1 plays critical roles in mitosis by regulating the spindle length, pole formation, and known for clustering extra centrosomes in cancer cells. Centrosome clustering is associated with the survival of cancer cells, but this phenomenon remains obscure in prostate cancer (PCa). The present study demonstrated that PCa cells showed centrosome amplification and clustering during interphase and mitosis, respectively. KIFC1 is highly expressed in PCa cells and tumor tissues of prostatic adenocarcinoma (PAC) patients. Up-regulation of KIFC1 facilitated the PCa cell survival in vitro by ensuring bipolar mitosis through clustering the multiple centrosomes, suggesting centrosome clustering could be a leading cause of prostate carcinogenesis. Conversely, the silencing of KIFC1 resulted in normal centrosome number or multipolar mitosis by inhibiting the clustering of amplified centrosomes in PCa cells. Besides, knockdown of KIFC1 by RNAi in PCa cells reduced cancer cell survival, and proliferation. KIFC1 interacted with centrosome structural protein Centrin 2 in clustering of amplified centrosomes in PCa cells to ensure the bipolar mitotic spindle formation. Knockdown of Centrin 2 in PCa cells inhibited the centrosome amplification and clustering. Moreover, up-regulated KIFC1 promotes PCa cell proliferation via progression of cell cycle possibly through aberrant activation of cyclin dependent kinase 1(Cdk1). Therefore, KIFC1 can be a prognostic marker and therapeutic target of PCa for inhibiting the cancer cell proliferation.     

Metamodel-based global optimization using fuzzy clustering for design space reduction

High fidelity analysis are utilized in modern engineering design optimization problems which involve expensive black-box models.For computation-intensive engineering design problems,efficient global optimization methods must be developed to relieve the computational burden.A new metamodel-based global optimization method using fuzzy clustering for design space reduction(MGO-FCR) is presented.The uniformly distributed initial sample points are generated by Latin hypercube design to construct the radial basis function metamodel,whose accuracy is improved with increasing number of sample points gradually.Fuzzy c-mean method and Gath-Geva clustering method are applied to divide the design space into several small interesting cluster spaces for low and high dimensional problems respectively.Modeling efficiency and accuracy are directly related to the design space,so unconcerned spaces are eliminated by the proposed reduction principle and two pseudo reduction algorithms.The reduction principle is developed to determine whether the current design space should be reduced and which space is eliminated.The first pseudo reduction algorithm improves the speed of clustering,while the second pseudo reduction algorithm ensures the design space to be reduced.Through several numerical benchmark functions,comparative studies with adaptive response surface method,approximated unimodal region elimination method and mode-pursuing sampling are carried out.The optimization results reveal that this method captures the real global optimum for all the numerical benchmark functions.And the number of function evaluations show that the efficiency of this method is favorable especially for high dimensional problems.Based on this global design optimization method,a design optimization of a lifting surface in high speed flow is carried out and this method saves about 10 h compared with genetic algorithms.This method possesses favorable performance on efficiency,robustness and capability of global convergence and gives a new optimization strat     

Design of cellular manufacturing systems using objective functional clustering algorithms

Cellular manufacturing (CM) has emerged as an alternative to conventional batch-type manufacturing owing to the former's capability of reducing set-up times, in-process inventories and throughput times. It provides the basis for implementation of just-in-time (JIT) and flexible manufacturing systems (FMS). The machine-part group formation is an important issue in the design of CMSs. This paper presents objective functional clustering algorithms for cell formation problems in the design of cellular manufacturing systems. A deterministic objective functional algorithm (hard clustering) and a fuzzy objective functional algorithm (fuzzy clustering) are used to form the part families and machine cells simultaneously. A collection of data sets from open literature is used to test these algorithms. A software package has been developed to verify the implementation.     

数据预处理对毫米波全息图像目标检测准确度的影响

毫米波图像被广泛应用于人体安检的违禁物品检测,但是现阶段不同的预处理方法对毫米波图像目标检测的准确度有着较大的影响。通过空间滤波、空间滤波加最大值聚类和空间滤波加最小值聚类的方法对数据进行预处理,比对不同方法目标检测的准确度。实验结果表明,以上3种方法均能有效提高目标检测的定位准确度,其中空间滤波为最佳预处理方法,准确率能达到92.3%,相较于未经预处理原始数据的准确度提高了约4%。因此基于空间位置的滤波方法比基于反射强度的聚类滤波方法能更有效地提高毫米波图像检测的准确度,为提升主动式毫米波全息图像目标检测准确度的任务提供了一种参考依据。     

An ant-based clustering algorithm for manufacturing cell design

This paper proposes a new part clustering algorithm that uses the concept of ant-based clustering in order to resolve machine cell formation problems. The three-phase algorithm mainly utilizes distributed agents which mimic the way real ants collect similar objects to form meaningful piles. In the first phase, an ant-based clustering model is adopted to form the initial part families. For the purpose of part clustering, a part similarity coefficient is modified and used in the similarity density function of the model. In the second phase, the K-means method is employed in order to achieve a better grouping result. In the third phase, artificial ants are used again to merge the small, refined part families into larger part families in a hierarchical manner. This would increase the flexibility of determining the number of final part families for the factory layout designer. The proposed algorithm has been developed into a software system called the ant-based part clustering system (APCS). In addition to part family formation, APCS performs the tasks of machine assignment and performance evaluation. Finally, performance evaluation of the proposed algorithm was conducted by testing some well-known problems from literature. The evaluation results show that the algorithm is able to solve the cell formation problems effectively.     

一种K-均值脸谱图聚类新算法

Chernoff脸谱图简单,类似卡通画,能图形化地表示多元数据。但脸谱图聚类算法具有主观性的巨大的对比工作量,脸谱特征分配困难。因此,本文提出一种新的脸谱图聚类算法,它合并了K均值聚类或模糊G均值聚类算法。IRIS和蔬菜油数据集的实验结果表明新算法优于传统的聚类算法。     

Grouping SMT PCB Assembly using FCM Algorithm

１IntroductionOneofthecharacteristicsoftheelectronicindustriesinHongKongisitsflexiblemanufacturingenvironment．Thetypeofproductionistypicallyalowvolume熏highmixtype．Lowvolumemeanstheproductionforeachbatchofproductionisrathersmall熏typically５００－５０００unitsperbatch．HighmixmeansthevarietyoftheproductionislargeandmanydifferenttypesofPCBsareproducedatthesametime．Insuchsituation熏thechangeoverfromonetypeofPCBtoanothertypeisquitefrequentandun鄄avoidable．ButthetimewastedinchangingPCBty…     

Data-driven modeling and predictive control for boiler–turbine unit using fuzzy clustering and subspace methods

This paper develops a novel data-driven fuzzy modeling strategy and predictive controller for boiler–turbine unit using fuzzy clustering and subspace identification (SID) methods. To deal with the nonlinear behavior of boiler–turbine unit, fuzzy clustering is used to provide an appropriate division of the operation region and develop the structure of the fuzzy model. Then by combining the input data with the corresponding fuzzy membership functions, the SID method is extended to extract the local state-space model parameters. Owing to the advantages of the both methods, the resulting fuzzy model can represent the boiler–turbine unit very closely, and a fuzzy model predictive controller is designed based on this model. As an alternative approach, a direct data-driven fuzzy predictive control is also developed following the same clustering and subspace methods, where intermediate subspace matrices developed during the identification procedure are utilized directly as the predictor. Simulation results show the advantages and effectiveness of the proposed approach.     

Phenomenological and algorithmic methods for the solution of inverse problems of electromagnetic testing

Modern methods for the solution of inverse problems of nondestructive testing are described. The discussion is mainly focused on the methods based on a mathematical model of the respective physical phenomenon (so-called phenomenological methods) and the methods based on the algorithms for the analysis of digital signals (so-called algorithmic methods). The phenomenological methods involving a mathematical model assume that the configuration of the flaws in a tested specimen is varied until the norm of the mismatch between the model solution and the experimentally obtained signal is minimized. A good result is only guaranteed if the physics of the phenomenon in the model is close to reality. In algorithmic methods, the inversion procedure applied to experimental data is considered as an image-recognition problem. In this case, the signal is identified as a representative of the classes associated with known types of flaws. The classification algorithms, which are most frequently used for electromagnetic testing, are developed through identification of diagnostic signatures. This approach assumes the use of an artificial neural network trained with the signals from a predefined database that corresponds to a broad variety of flaws.