Process Control Issues Research Articles

In-Line Process and Product Control Using Spectroscopy and Multivariate Calibration

The quality of a product is dynamic in nature and develops over time. We present a case study from the food industry in which the concept of measuring the end-product quality is extended to incorporate the shelf-life period of the product with practical examples showing the harm of ignoring the changes that occur in the product quality as a function of time. The article also addresses the use of in-line spectroscopy to relate the variations in the input parameters, such as the raw materials, and the process variables to the final product quality over the entire shelf-life of the product. We also discuss multivariate statistical process control and monitoring issues.

Stability analysis of reactive sputtering process with variable sticking coefficients

In reactive sputtering, the introduction of reactive gas would create a hysteresis transition from metal to compound mode in both the target and the substrate. The hysteresis transition is characterized by a sudden change in partial pressure, sputtering rate and fraction of compound formation, etc. Therefore, the stability is an important issue of process control. In this paper, a mathematical model with variable sticking coefficients based on surface kinetics is introduced to study the process stability. The variable sticking coefficient represents different mechanisms for surface reactions from the type of Langmuir to precursors. To facilitate the analysis, several nondimensional parameters are introduced and used for formulation. Results show that, when the chemical reaction on a substrate is moderate, a higher sputtering yield of the compound leads to a more stable steady-state at lower inflow rates. However, when the ratio of sputtering yield (compound/metal) is zero, there is no hysteresis transition because of the deposition of pure metal (single phase) on the substrate. For different sticking mechanisms, the precursor type is found to make the surface easier to saturate with compound due to its higher default sticking coefficient and the lower pressure and inflow rate for the hysteresis transition.

Applying State Space to SPC: Monitoring Multivariate Time Series

Monitoring cross-sectional and serially interdependent processes has become a new issue in statistical process control (SPC). In up-to-date SPC literature, Kalman filtering was reported to monitor univariate autocorrelated processes. This paper applies a Kalman filter or state-space method for SPC to monitoring multivariate time series. We use Aoki's approach to estimate the parameter matrices of a state-space model. Multivariate Hotelling T 2 control charts are employed to monitor the residuals of the state-space. Examples of this approach are illustrated.

OPTIMIZING FEEDFORWARD NEURAL NETWORKS FOR CONTROL CHART PATTERN RECOGNITION THROUGH GENETIC ALGORITHMS

Pattern recognition is an important issue in statistical process control (SPC) because unnatural patterns exhibited by control charts can be associated with specific assignable causes adversely affecting the process. Artificial neural networks have been widely investigated as an effective approach to control chart pattern (CCP) recognition in recent years. However, an overwhelming majority of these applications has used trial-and-error experiments to determine the network architecture and training parameters, which are crucial to the performance of the network. In this paper, the genetic algorithm (GA) is used to evolve the configuration and the training parameter set of the neural network to solve the online CCP recognition problem. Numerical results are provided that indicate that the proposed GA can evolve neural network architecture while simultaneously determining training parameters to maximize efficiently the performance of the online CCP recognizers. Because the population size is a major parameter of GA processing speed, an investigation was also conducted to identify the effects of the population size on the performance of the proposed GA. This research further confirms the feasibility of using GA to evolve neural networks. Although a back-propagation-based CCP recognizer is the particular application presented here, the proposed GA methodology can be applied to neural networks in general.

Stability analysis of reactive sputtering process

In reactive sputtering, the introduction of reactive gas would create a hysteresis transition from metal to compounds (oxides, nitrides and carbides, etc.) in both target and substrate. The hysteresis transition is characterized by a sudden change in partial pressure, cathode voltage, sputtering rate and fraction of compound formation. Therefore, the stability is an important issue of process control. In this article, a simple mathematical model based on Berg's original proposal is used to study the stability of steady state solutions. In order to facilitate the analysis, several non-dimensional parameters are identified and used for formulation. The stability is then investigated in two folds. For a small variation away from the steady state, the linear perturbation method is utilized to reduce the model to an eigenvalue problem. For large deviation from the steady state, a full numerical analysis is employed to obtain the steady state. Results show that an unsteady system converges to steady states relatively fast at low inflow rates and the positive range of eigenvalue are coherent with the presence of hysteresis loop. With the aid of the eigenvalue, it is also found that when the chemical reaction on the substrate is moderate, a higher sputter yield of the compound leads to a more stable steady state at lower inflow rates. Finally, a convergence parameter similar to the Lyapunov number is introduced for full numerical analysis to give a qualitative measure for the stability at steady state.

A Generalized Least Square Model for the Determination of Monomer Reactivity Ratios in Free Radical Copolymerization Systems

AbstractThe reliable estimation of monomer reactivity ratios is a key problem in the development of process models, and, thus, is an important issue in both process design and control. Copolymers of isobutyl methacrylate (i‐BMA), and lauryl methacrylate (LMA) were prepared by free radical bulk copolymerization at 70 °C using 2,2′–azobisisobutyronitrile (AIBN) as initiator. The monomer reactivity ratios as well as the 95% confidence intervals were calculated from Fineman‐Ross, Ezrielev‐Brokhina‐Roskin, Kelen & Tudos, extended Kelen & Tudos and a linear iterative method proposed by Mao and Huglin. The estimation process was performed by applying nonlinear techniques based on Marquardt optimization, “ordinary least square” (OLS) and “generalized least square” (GLS) approaches and the results were compared with each other. The novel GLS estimator program developed during this study was employed to ascertain the shape of the copolymer composition curve and good agreement between experimental and calculated copolymer composition was obtained. Also, both real and simulated data was presented, which show GLS method is a statistically valid technique for estimating reactivity ratios by accounting response error structure. This new estimation method handles regression models with error terms that are heteroskedastic or serially correlated, or both.95% joint confidence interval for reactivity ratios from “NLLS” analysis.magnified image95% joint confidence interval for reactivity ratios from “NLLS” analysis.

Business Process Design: Correlates of Success and Failure

This article summarizes 75 organizational change efforts that had business process design (BPD) as an objective. BPD was found to be a common type of organizational change and usually occurred in combination with other types of change. The success rate for process design was low. Success was more likely when the sponsors were perceived to be mid-level rather than senior executives. BPD was most often undertaken because of cost trends, product problems, process control issues, customer complaints and suggestions, and competition. Statistical data were most often cited to describe successful process design, while unsuccessful change efforts were usually described by opinions. Success correlated most highly with the variables that reflected employee needs, leadership, and project management. Failure correlated most strongly with ineffective leadership and the clash with the existing culture. Success factors and barriers for cultural change resembled the profile for other types of organizational change. The author derived seven implications for managing process design and implementation and discussed difficulties in comparing this study with other research efforts to identify factors that facilitate or inhibit process design.

Dealing with missing data in MSPC: several methods, different interpretations, some examples

AbstractThis paper addresses the problem of using future multivariate observations with missing data to estimate latent variable scores from an existing principal component analysis (PCA) model. This is a critical issue in multivariate statistical process control (MSPC) schemes where the process is continuously interrogated based on an underlying PCA model. We present several methods for estimating the scores of new individuals with missing data: a so‐called trimmed score method (TRI), a single‐component projection method (SCP), a method of projection to the model plane (PMP), a method based on the iterative imputation of missing data, a method based on the minimization of the squared prediction error (SPE), a conditional mean replacement method (CMR) and various least squared‐based methods: one based on a regression on known data (KDR) and the other based on a regression on trimmed scores (TSR). The basis for each method and the expressions for the score estimators, their covariance matrices and the estimation errors are developed. Some of the methods discussed have already been proposed in the literature (SCP, PMP and CMR), some are original (TRI and TSR) and others are shown to be equivalent to methods already developed by other authors: iterative imputation and SPE methods are equivalent to PMP; KDR is equivalent to CMR. These methods can be seen as different ways to impute values for the missing variables. The efficiency of the methods is studied through simulations based on an industrial data set. The KDR method is shown to be statistically superior to the other methods, except the TSR method in which the matrix to be inverted is of a much smaller size. Copyright © 2002 John Wiley & Sons, Ltd.

Development of process control in sheet metal forming

In sheet metal forming processes, the blank holder force controls the material flow into the die cavity, which is critical to producing a good part. Process control can be used to adjust the blank holder force in-process based on tracking a reference punch force trajectory to improve part quality and consistency. Key issues in process control include process controller and reference punch force trajectory design. The purpose of this paper is to present a systematic approach to the design and implementation of a suitable process controller and an optimal reference punch force trajectory. The approach includes modeling of the sheet metal forming process, design of the process controller, and determination of the optimal punch force trajectory. Experimental results from U-channel forming show that a suitable process controller can be designed through simulation and an optimal reference punch force trajectory can be synthesized through experiments. The proposed development should be useful in designing and implementing process control in sheet metal forming processes.

Robustness of the neural network based control chart pattern recognition system to non‐normality

Control chart pattern recognition is a critical issue in statistical process control, as unnatural patterns on control charts are often associated with specific assignable causes adversely affecting the process. Several researchers have recently applied neural networks to pattern recognition for control charts. However, nearly all studies in this area assume that the in‐control process data in the control charts follow a normal distribution. This assumption contradicts the facts of practical manufacturing situations. This paper investigates how non‐normality affects the performance of neural network based control chart pattern recognition models. Extensive performance evaluation was carried out using simulated data with various non‐normalities. The non‐normality was measured in skewness and kurtosis. Numerical results indicate that the neural network based control chart pattern recognition models still perform well in a non‐normal distribution environment in terms of recognition accuracy and speed.

A Performance Evaluation of Lightweight Non-Contact Pavement Profilers

Abstract Pavement smoothness was recently identified as the most significant factor the motoring public uses to judge the quality of U.S. roadways. A new generation of lightweight profilers has been developed that have the potential of providing nearly instantaneous smoothness measurements, giving contractors a tool to identify and address process control issues promptly and cost effectively. This new technology was cooperatively evaluated by a team of Purdue University and Indiana Department of Transportation (INDOT) researchers to assess the performance of the devices, as well as the potential for their use in Indiana. A successful field test was conducted using four lightweight profilers that performed five replicate measurements at each of six sites: three, hot mix asphalt (HMA) and three, portland cement concrete (PCC). Smoothness data were reported in terms of International Roughness Index (IRI) and Profile Index with 0.0 and 0.2 in. (PI-0.0 and PI-0.2) blanking bands. The field test data showed good repeatability and poor reproducibility in terms of the smoothness indices considered, the profiles generated, and the identification of must-grind locations. Estimates of precision for the devices were determined in accordance with ASTM standards. The observed precision statements as well as statistical tests of between-vendor performance suggest that the performance of the devices, especially the reproducibility, should be improved prior to implementation of their use in construction specifications.

Small Sample Properties of an Adaptive Filter Applied to Low Volume SPC

In many manufacturing environments, such as the nuclear weapons complex, emphasis has shifted from the regular production and delivery of large orders to infrequent small orders. However, the challenge to maintain the same high quality and reliability standards while building much smaller lot sizes remains. To meet this challenge, specific areas need more attention, including fast and on-target process start-up, low volume statistical process control, process characterization with small experiments, and estimating reliability given few actual performance tests of the product. In this paper we address the issue of low volume statistical process control. We investigate an adaptive filtering approach to process monitoring with a relatively short time series of autocorrelated data. The emphasis is on estimation and minimization of mean squared error rather than the traditional hypothesis testing and run length analyses associated with process control charting. We develop an adaptive filtering technique that assumes initial process parameters are unknown, and updates the parameters as more data become available. Using simulation techniques, we study the data requirements (the length of a time series of autocorrelated data) necessary to adequately estimate process parameters. We show that far fewer data values are needed than is typically recommended for process control applications. We also demonstrate the techniques with a case study from the nuclear weapons manufacturing complex.

Evaluation of the Run-Length Probability Distribution for CUSUM Charts: Assessing Chart Performance

This article provides a fast and accurate algorithm to compute the run-length probability distribution for cumulative sum charts to control process mean. This algorithm uses a fast and numerically stable recursive formula based on accurate Gaussian quadrature rules throughout the whole range of the computed run-length distribution and, therefore, improves the numerical efficiency and accuracy of existing methods. The algorithm may detect whether or not the geometric approximation is adequate and, when it is possible, it allows switching to the geometric recursion. The procedure may be applied not only to the normal distribution but also to nonsymmetric and long-tailed continuous distributions, some examples of which are provided. Methods to assess chart performance according to the run-length distribution, as well as some multivariate issues in statistical process control, are considered.

Statistics in Preclinical Pharmaceutical Research and Development

eds. H.-J. Lenz and P.-Th. Wilrich, Heidelberg, Germany: SpringerVeriag, 208-228. Stoumbos, Z. G., and Reynolds, M. R., Jr. (1997), “Control Charts Applying a Sequential Test at Fixed Sampling Intervals,” Journal of Quality Technology, 29, 2140. (2000a), “The SPRT Control Chart for the Process Mean With Samples Starting at Fixed Times,” Nonlinear Analysis, (to appear). (2000b), “Robustness to Non-Normality and Autocorrelation of Individual Control Charts,” Journal of Statistical Computation and Simulation, Special Issue on Statistics: Modeling of Processes in Science and Industry, 66, 145-187. Tagaras, G. (1998), “A Survey of Recent Developments in the Design of Adaptive Control Charts,” Journal of Quality Technology, 30, 21 2-23 1. Tsung, F., Shi, J., and Wu, C. F. J. (1999), “Joint Monitoring of PIDControlled Processes,” Journal of Quality Technology, 31, 275-285. Woodall, W. H. (1986), “Weaknesses of the Economic Design of Control Charts” (letter to the editor), Technometrics, 28, 408-410. Woodall, W. H., and Montgomery, D. C. (1999), “Research Issues and Ideas in Statistical Process Control,” Journal of Quality Technology, 31,

Steady-state operability characteristics of reactors

Operability analysis helps the process designer to consider process control issues quite early in the design process. This analysis should lead to the calculation of measure which quantifies the operability of a design. Search for this measure has been an active area of research in the last two decades. Recently, Vinson and Georgakis proposed a steady-state operability index (OI) involving both the servo and regulatory aspects. Their primary focus was on linear systems. This paper demonstrates the utility of the OI approach to chemical reactors. Two designs of a nonlinear CSTR, with and without multiple steady states, are examined.

Reliable nonlinear parameter estimation using interval analysis: error-in-variable approach

Parameter estimation is a key problem in the development of process models, both steady- and unsteady-state, and thus is an important issue in both process design and control. The error-in-variable (EIV) approach differs distinctly from the standard approach in that measurement errors in both dependent and independent system variables are taken into account when formulating the objective function in the parameter estimation problem. It is not uncommon for the objective function in nonlinear parameter estimation problems to have multiple local optima. However, the usual methods used to solve these problems are local methods that offer no guarantee that the global optimum, and thus the best set of model parameters, has been found. We demonstrate here a technique, based on interval analysis, that can solve the EIV parameter estimation problem with complete reliability, providing a mathematical and computational guarantee that the global optimum is found. As examples, we consider the estimation of parameters in both steady and unsteady-state models, including a vapor—liquid equilibrium (VLE) model, a CSTR model, and a reaction kinetics model.

Research Issues and Ideas in Statistical Process Control

An overview is given of current research on control charting methods for process monitoring and improvement. The discussion includes a historical perspective along with ideas for future research. Research topics include, for example, variable sample size and sampling interval methods, economic designs, attribute data methods, charts based on autocorrelated observations, multivariate methods, and nonparametric methods. Recommendations and references are provided to those interested in pursuing research ideas in statistical process control (SPC). Some issues regarding the relevance of SPC research are also discussed.

Recognition of control chart concurrent patterns using a neural network approach

Pattern recognition is an important issue in Statistical Process Control, as unnatural patterns exhibited by control charts can be associated with specific assignable causes adversely affecting the process. Neural network approaches to recognition of control chart patterns have been developed by several researchers in recent years, but to date these have been focused on recognition and analysis of single patterns such as sudden shifts, linear trends or cyclic patterns. This paper investigates the detection of concurrent patterns where more than one pattern exists simultaneously. The topology and training of a Back-Propagation Network (BPN) system is described. Extensive performance evaluation has been carried out using simulated data to develop a range of average run length-related performance indices, including new performance indices that are proposed to describe concurrent patterns recognition performance. Two evaluation scenarios were evaluated: in the first, unnatural patterns are already present; while in the second, patterns may appear progressively at any time. Numerical results are provided that indicate that the pattern recognizer can perform very well in the first scenario, while it performs effectively but not without deficiencies for some specific pattern combinations in the second evaluation approach. Limitations and potential improvements in the concurrent pattern recognition scheme are also discussed.

Keys to Lime Handling and Process Control

Part 1 (June 1998 issue of Opflow) of this series discussed several problems and associated solutions regarding handling and feeding lime to the treatment process. Once lime has been fed to the point of application, many process control issues can be addressed to optimize and control lime use. Part 2 will discuss some of the solutions that have been developed to enhance the performance of lime in various treatment applications.

Mathematical models of transport phenomena associated with arc-welding processes: a survey

A comprehensive survey is presented of research on transport phenomena, that is, heat flow, fluid flow, mass transfer, and electrodynamics, associated with arc welding. Both gas-metal arc-welding (GMAW) and gas-tungsten arc welding (GTAW) systems are considered, and appropriate emphasis is placed on the intrinsic differences between these systems. The topics discussed include the behaviour of the arc, and its interaction with both the electrode and the surrounding medium, on the one hand, and the behaviour of the weld pool on the other. The limited work done on the coupling between arcs, electrodes, and weld pools is also discussed, as are some process-control issues, but the problems of thermal stress evolution and stress-induced deformation in the solid state are not treated. A critical discussion of the comparison between model predictions and experimental measurements is a key feature of the presentation, which concludes with a discussion of the critical problems that still need to be addressed.