In most manufacturing processes, identifying the faulty process variables that may lead to process changes is crucial for quality engineers and practitioners. There are several parametric procedures for identifying fa...
详细信息
In most manufacturing processes, identifying the faulty process variables that may lead to process changes is crucial for quality engineers and practitioners. There are several parametric procedures for identifying faulty variables with the assumption that they follow multivariate normal distributions. However, in practice, the normality assumption restricts the applicability of such procedures in identifying the faulty variables. In addition, conventional procedures for fault identification are often computationally expensive, especially in high-dimensional processes. Therefore, this article proposes a data-driven Bayesian approach for fault identification that addresses the limitations posed by the normality assumption. The proposed approach is computationally efficient for high-dimensional data compared with existing approaches. Experimental results with various simulation studies and real-life data sets demonstrate the effectiveness of the proposed procedure.
In multivariate statistical process control (MSPC) applications, process mean shifts sometimes occur in only a few components. To solve this MSPC problem, many control charts were proposed in the literature. Most of t...
详细信息
In multivariate statistical process control (MSPC) applications, process mean shifts sometimes occur in only a few components. To solve this MSPC problem, many control charts were proposed in the literature. Most of these charts assumed that the multivariate quality characteristics are normally distributed. Among them, the control chart proposed by Zou and Qiu (2009), incorporating the least absolute shrinkage and selection operator (LASSO) method into the EWMA scheme, has the best overall performance. In this paper, we extend the classical multivariate LASSO control chart to a robust version that has an affine-invariance property and is distribution free under the family of elliptical direction distributions, indicating that the in-control run-length distribution is the same for any continuous distribution in this family and the control limit can be acquired from the multivariate standard normal distribution. Our simulation results show that the proposed method is very efficient in detecting various sparse shifts under heavy-tailed and skewed multivariate distributions. In addition, it is easy to implement with an iterative algorithm and the least angle regression (LARS) algorithm. White-wine data illustrates that the proposed control chart performs quite well in applications.
In this work, the integration of ARMA filters into the multivariate statistical process control (MSPC) framework is presented to improve the monitoring of large-scale industrial processes. As demonstrated in the paper...
详细信息
In this work, the integration of ARMA filters into the multivariate statistical process control (MSPC) framework is presented to improve the monitoring of large-scale industrial processes. As demonstrated in the paper, such filters can remove auto-correlation from the monitored variables to avoid the production of false alarms. This is exemplified by application studies to a synthetic example from the literature and to the Tennessee Eastman benchmark process. (C) 2004 Elsevier Ltd. All rights reserved.
multivariatestatistical methods for online process monitoring have been widely applied to chemical, biological, and engineered systems. While methods based on principal component analysis (PCA) are popular, more rece...
详细信息
multivariatestatistical methods for online process monitoring have been widely applied to chemical, biological, and engineered systems. While methods based on principal component analysis (PCA) are popular, more recently kernel PCA (KPCA) and locally linear embedding (LLE) have been utilized to better model nonlinear process data. Additionally, various forms of dynamic and adaptive monitoring schemes have been proposed to address time-varying features in these processes. In this analysis, we extend a common simulation study in order to account for autocorrelation and nonstationarity in process data and comprehensively compare the monitoring performances of static, dynamic, adaptive, and adaptive-dynamic versions of PCA, KPCA, and LLE. Furthermore, we evaluate a nonparametric method to set thresholds for monitoring statistics and compare results with the standard parametric approaches. We then apply these methods to real-world data collected from a decentralized wastewater treatment system during normal and abnormal operations. From the simulation study, adaptive-dynamic versions of all three methods generally improve results when the process is autocorrelated and nonstationary. In the case study, adaptive-dynamic versions of PCA, KPCA, and LLE all flag a strong system fault, but nonparametric thresholds considerably reduce the number of false alarms for all three methods under normal operating conditions.
A control procedure is presented for monitoring changes in variation for a multivariate normal process in a Phase II operation where the subgroup size, m, is less than p, the number of variates. The methodology is bas...
详细信息
A control procedure is presented for monitoring changes in variation for a multivariate normal process in a Phase II operation where the subgroup size, m, is less than p, the number of variates. The methodology is based on a form of Wilk' statistic, which can be expressed as a function of the ratio of the determinants of two separate estimates of the covariance matrix. One estimate is based on the historical data set from Phase I and the other is based on an augmented data set including new data obtained in Phase II. The proposed statistic is shown to be distributed as the product of independent beta distributions that can be approximated using either a chi-square or F-distribution. An ARL study of the statistic is presented for a range of conditions for the population covariance matrix. Cases are considered where a p-variate process is being monitored using a sample of m observations per subgroup and mp. Data from an industrial multivariateprocess is used to illustrate the proposed technique.
Conventional multivariatecontrol charts usually focus on a specific process shifts range (small or large), and they cannot get the knowledge of manufacturing process through the learning of in-control data and be eff...
详细信息
Conventional multivariatecontrol charts usually focus on a specific process shifts range (small or large), and they cannot get the knowledge of manufacturing process through the learning of in-control data and be effective over the whole range of mean shifts, due to the characteristics of their own structures. In this paper, an effective combined multivariatecontrol chart (named CDD chart, i.e., combined D-MCUSUM and D chart) with an adaptive control limit is proposed to improve the overall detection ability of monitoring techniques in multivariate statistical process control. Besides, this paper also provides a basic methodology for designing the adaptive control limit and recommended values of some key parameters (e.g. window size) for a better application. Based on these, a bivariate simulation experiment is conducted to evaluate the performance of the proposed control chart. Simulation results show that the CDD chart offers a better overall performance, compared with regular control charts (e.g. MCUSUM). In addition, a real industrial case also illustrates the effectiveness of the proposed control chart in applications.
A 2(3)-factorial experiment was carried out in an industrial plant producing biofuel pellets with sawdust as feedstock. The aim was to use on-line near infrared (NIR) spectra from sawdust for real time predictions of ...
详细信息
A 2(3)-factorial experiment was carried out in an industrial plant producing biofuel pellets with sawdust as feedstock. The aim was to use on-line near infrared (NIR) spectra from sawdust for real time predictions of moisture content, blends of sawdust and energy consumption of the pellet press. The factors varied were: drying temperature and wood powder dryness in binary blends of sawdust from Norway spruce and Scots pine. The main results were excellent NIR calibration models for on-line prediction of moisture content and binary blends of sawdust from the two species, but also for the novel finding that the consumption of electrical energy per unit pelletized biomass can be predicted by NIR reflectance spectra from sawdust entering the pellet press. This power consumption model, explaining 91.0% of the variation, indicated that NIR data contained information of the compression and friction properties of the biomass feedstock. The moisture content model was validated using a running NIR calibration model in the pellet plant. It is shown that the adjusted prediction error was 0.41% Moisture content for grinded sawdust dried to ca. 6-12% moisture content. Further, although used drying temperatures influenced NIR spectra the models for drying temperature resulted in low prediction accuracy. The results show that on-line NIR can be used as an important tool in the monitoring and control of the pelletizing process and that the use of NIR technique in fuel pellet production has possibilities to better meet customer specifications, and therefore create added production values. (C) 2008 Elsevier Ltd. All rights reserved.
statisticalprocesscontrol (SPC) techniques that originated in manufacturing have also been used to monitoring the quality of various service processes, which can be characterized by one or several variables. In the ...
详细信息
statisticalprocesscontrol (SPC) techniques that originated in manufacturing have also been used to monitoring the quality of various service processes, which can be characterized by one or several variables. In the literature, these variables are usually assumed to be either continuous or categorical. However, in reality, the quality characteristics of a service process may include both continuous and categorical variables (i.e., mixed-type variables). Direct application of conventional SPC techniques to monitor such mixed-type variables may cause increased false alarm rates and misleading conclusions. One promising solution is the kernel distancebased chart (K-chart), which makes use of Support Vector Machine (SVM) methods and requires no assumption on the variable distribution. This article provides an improved design of the SVM-based K-chart. A systematic approach to parameter selection for the considered charts is provided. An illustration and comparison are presented based on a real example from a logistics firm. The results confirm the improved performance obtained by using the proposed design scheme.
Over the past decade, multivariate statistical process control (MSPC) methods have been proven, in the process industries, to be an effective tool for process monitoring, modelling and fault detection. This paper desc...
详细信息
Over the past decade, multivariate statistical process control (MSPC) methods have been proven, in the process industries, to be an effective tool for process monitoring, modelling and fault detection. This paper describes the development of a real-time monitoring solution for a complex petroleum refining process with an installed multivariable model predictive controller. The developed solution was designed to track the time-varying and non-stationary dynamics of the process and for improved isolation capabilities, a multiblock approach was applied. The paper highlights the systematic and generic approach that was followed to develop the monitoring solution and stresses the importance of exploiting the knowledge of experienced plant personnel when developing any such system. (c) 2007 Elsevier Ltd. All rights reserved.
multivariate statistical process control is used for simultaneously monitoring several process variables. The original artificial contrasts (AC) are very useful for monitoring inhomogeneously distributed data with an ...
详细信息
multivariate statistical process control is used for simultaneously monitoring several process variables. The original artificial contrasts (AC) are very useful for monitoring inhomogeneously distributed data with an indicator variable. The cluster-based AC improve it by considering separated clusters, respectively. Then the artificial data used for the AC overlap each cluster. Numerical experiments show that our method outperforms existing methods in terms of Type-II error rate.
暂无评论