Guaranteed in‐control performance of the synthetic chart with estimated parameters

Control charts are usually investigated under the assumption of known process parameters. In practice, however, the process parameters are rarely known and they have to be estimated from different Phase I data sets. The properties of control charts with estimated parameters are usually investigated with the unconditional average of the average run length. Control chart's performance is known to vary among practitioners because of the use of different Phase I data sets. Considering the between‐practitioners variability in control chart's performance, the standard deviation of the average run length is developed to reevaluate the properties of the synthetic chart with estimated parameters. Because of the limited amount of Phase I data in practice, the bootstrap method is used as a good adjustment technique for the synthetic chart's parameters.     

Optimal design of the side‐sensitive modified group runs (SSMGR) chart when process parameters are estimated

A side‐sensitive modified group runs (SSMGR) chart is commonly examined under the assumption of known process parameters (denoted as Case‐K). However, in practical situations, the process parameters are seldom known, and it is essential to estimate them from in‐control (IC) phase I samples. From the study of the minimum amount of phase I samples, m that is required by the SSMGR chart when the process parameters are estimated (represented as Case‐U), to achieve practically the same performance as that of the Case‐K, it was found that a huge number of IC phase I samples are needed. Due to time and cost factors, accumulating a substantial amount of IC phase I samples is difficult. To circumvent this problem, we study the optimal designs of the Case‐U SSMGR chart, based on both the zero and steady states' average number of observations to signal (ANOS) and expected ANOS (EANOS). Tables containing optimal charting parameters are provided to ease the implementation of the Case‐U SSMGR chart. By using these optimal parameters specifically designed for the Case‐U SSMGR chart, we can obtain similar zero and steady states' ANOS and EANOS performances to that of the Case‐K. The use of the optimal Case‐U SSMGR chart is demonstrated with real data taken from a hard‐bake process in a manufacturing company.     

Economic Design of the Variable Parameters Control Chart with a Corrected A&L Switching Rule

The variable parameters (VP) control chart is more efficient than the fixed parameters control chart. However, an obvious shortcoming of the VP control chart is its excessive number of switches among different design parameters, which significantly increases the complexity of operations for administer and the costs of maintaining the control chart. Amin and Letsinger (1991) proposed a switching rule, denoted by A&L switching rule, to reduce the number of switches. In this paper, we investigated the economic design of the VP control chart by using the corrected A&L switching rule scheme, which makes some corrections to the existing study of the A&L switching rule scheme. A three‐state Markov model is proposed to derive the performance measures of the suggested control chart. An expected cost model for the process is established with the cost function derived. The genetic algorithm is then employed to search for the solution for the economic design of the proposed control chart. The results obtained show that the developed model can significantly reduce the average number of the parameter's switches and save the expected cost in comparison with the conventional VP control scheme in detecting small and moderate mean shifts. A sensitivity analysis is also carried out to examine the effects of cost and model parameters on the solution of the economic design for the proposed control chart. The analysis demonstrates the expected cost per time unit is positively affected by the cost associated with a false signal, the average cost of repairs when a true signal is identified, the sampling cost and the cost associated with a switch, and is negatively affected by the cost associated with running the process in‐control per hour. Copyright © 2013 John Wiley & Sons, Ltd.     

The Variable Sample Size Chart with Estimated Parameters

The VSS chart, dedicated to the detection of small to moderate mean shifts in the process, has been investigated by several researchers under the assumption of known process parameters. In practice, the process parameters are rarely known and are usually estimated from an in‐control Phase I data set. In this paper, we evaluate the (run length) performances of the VSS chart when the process parameters are estimated, we compare them in the case where the process parameters are assumed known and we propose specific optimal control chart parameters taking the number of Phase I samples into account.     

Economically Optimum Design of a Synthetic Chart

This paper proposes an economic model for the synthetic chart. The synthetic chart is an integration of the chart and the CRL chart. A simplified algorithm to obtain the optimal parameters of the synthetic chart which minimizes the expected cost function is introduced. Numerical examples based on different values of input parameters are given, and sensitivity analyses of the parameters are performed. The input parameters which have a significant impact on the cost and choice of optimal parameters of the synthetic chart are identified. The effect of incorrect estimation of the input parameters is also investigated, and it is found that the optimal design parameters are quite robust to changes in the input parameters, except the shift parameter. It is also shown that if the chart cannot be operated at the economically optimal level, there is still a large choice of parameters to choose from which does not result in a large increase in cost. All the above analyses and results are based on numerical examples and verified through simulation over a wide range of parameter values. Comparisons are also performed among the synthetic Shewhart and EWMA charts. Based on the numerical examples and simulation over a wide range of parameter values, it is shown that the synthetic chart has better economic performance than the other two control charts under most conditions. Copyright © 2011 John Wiley & Sons, Ltd.     

Run sum X¯ control chart with estimated process parameters

The run sum $\overline{X}$ control chart is usually investigated under the assumption of known process parameters. In practice, process parameters are rarely known and they need to be estimated from an in‐control Phase I dataset. However, different practitioners use different numbers of Phase I samples to estimate the process parameters. As a result, the commonly used performance measure, ie, the average run length becomes a random variable. In this study, we present a run sum $\overline{X}$ control chart with estimated process parameters and use the standard deviation of the average run length to evaluate the average of the average run length performance of the run sum $\overline{X}$ chart when process parameters are estimated. Based on the standard deviation of the average run length criterion, the number of Phase I samples required by the estimated process parameter–based run sum $\overline{X}$ chart to have an average of the average run length performance close to that obtained under the assumption of known process parameters is recommended.     

Re‐evaluation of the VSI‐ chart performance with estimated parameters

The performance of the variable sampling interval‐ (VSI‐ ) chart with estimated parameters has been investigated on the basis of the average time to signal (ATS) and standard deviation of time to signal (SDTS) in past research studies. Since the values of ATS and SDTS vary from practitioner to practitioner, the use of these 2 measures is not reliable. The use of different historical data sets in phase I results in varying parameter estimates, control limits, warning limits, ATS, and SDTS values. In this study, we use the standard deviation of average time to signal (SDATS) to evaluate and compare the performance of the VSI‐ chart with known parameters and estimated parameters. This study shows that variation reduction in ATS values requires a larger than previously recommended phase I data. Also, detection of up to moderate shifts in the process mean with the desired ATS value would be achievable with the number of samples recommended in the past, but the in‐control performance of the chart would not be reliable. Furthermore, we evaluate the effect of using large and small desired values of ATS₀ on the performance of in‐control and out‐of‐control VSI‐ chart. We also study the effects of estimating the mean and standard deviation on the ATS values using numerical simulation. Finally, we present a method based on warning and control limits coefficients for the estimated parameters case to reduce the number of samples required in phase I.     

Effectiveness of phase I applications for identifying randomly scattered out‐of‐control observations and estimating control chart parameters

Estimation of unknown process parameters with fixed‐size samples are studied in the following. The standard textbook approach for phase I control chart implementation with a Shewhart control chart is evaluated for the case of normally distributed independent observations with random sampling. The charts are simultaneously implemented by generating observations that have a given percentage of randomly scattered out‐of‐control observations. Simulating the phase I steps, where out‐of‐control samples are detected iteratively by determining trial control limits, identifying samples exceeding these limits, and revising the control limits, the standard practice is evaluated in terms of both detection performance and quality of parameter estimates. It is shown that standard phase I control chart implementations with 3‐σ‐limits may perform very poorly in identifying true out‐of‐control observations and providing a reference set of in‐control observations for estimation in some practical settings. A chart design with 2‐σ‐limits is recommended for a successful phase I analysis.     

Process Yield for Multivariate Linear Profiles with One‐sided Specification Limits

The new investigation of profile monitoring is focused mainly on a process with multiple quality characteristics. Process yield has been used widely in the manufacturing industry, as an index for measuring process capability. In this study, we present two indices and to measure the process capability for multivariate linear profiles with one‐sided specification limits under mutually independent normality. Additionally, two indices and are proposed to measure the process capability for multivariate linear profiles with one‐sided specification limits under multivariate normality. These indices can provide an exact measure of the process yield. The approximate normal distributions for and are constructed. A simulation study is conducted to assess the performance of the proposed approach. The simulation results show that the estimated value of performs better as the number of profiles increases. Two illustrative examples are used to demonstrate the applicability of the proposed approach. Copyright © 2015 John Wiley & Sons, Ltd.     

VSSI median control chart with estimated parameters and measurement errors

An adaptive control chart called Shewhart chart with variable sample size and sampling interval (VSSI) is quicker than Shewhart chart, chart with variable sample size (VSS), and chart with variable sampling interval (VSI) in detecting the mean shifts of a normal process. In practice, the effects of measurement errors on control charts should be included. In this study, we present an VSSI median control chart with estimated parameters in the presence of measurement errors for a normal process. The average time to signal (ATS) is computed by using the Markov chain approach. The results show that the VSSI median control chart performs better than the Shewhart median, VSS median, and VSI median control charts in terms of ATS. The design parameters of the proposed chart are provided. Two examples are used to illustrate the application of the proposed control chart.     

Adaptive multivariate double sampling and variable sampling interval Hotelling's T2 charts

In this article, two adaptive multivariate charts, which combine the double sampling (DS) and variable sampling interval (VSI) features, called the adaptive multivariate double sampling variable sampling interval T² (AMDSVSI T²) and the adaptive multivariate double sampling variable sampling interval combined T² (AMDSVSIC T²) charts, are proposed. The real purpose of using the proposed charts is to provide flexibility by enabling the sampling interval length of the DS T² chart to be varied so that the chart's sensitivity can be enhanced. The fundamental difference between the two proposed charts is that when a second sample is taken, the AMDSVSI T² chart uses the information of the combined sample mean vectors while the AMDSVSIC T² chart uses the information of the combined T² statistics, in deciding about the process status. This research is motivated by existing combined DS and VSI charts in the literature, which show convincing performance improvement over the standard DS chart. Consequently, it is believed that adopting this existing approach in the multivariate case will enable superior multivariate DS charts to be proposed. Numerical results show that the proposed charts outperform the existing standard T² and other adaptive multivariate charts, in detecting shifts in the mean vector, for the zero‐state and steady‐state cases. The performances of both charts when the shift sizes in the mean vector are unknown are also measured. The application of the AMDSVSI T² chart is illustrated with an example.     

On the Performance of Shewhart‐type Synthetic and Runs‐rules Charts Combined with an Chart

A synthetic chart is a combination of a conforming run‐length chart and an chart, or equivalently, a 2‐of‐(H + 1) runs‐rules (RR) chart with a head‐start feature. However, a synthetic chart combined with an chart is called a Synthetic‐ chart. In this article, we build a framework for Shewhart Synthetic‐ and improved RR (i.e., 1‐of‐1 or 2‐of‐(H + 1) without head‐start) charts by conducting an in‐depth zero‐state and steady‐state study to gain insight into the design of different classes of these schemes and their average run‐length performance using the Markov chain imbedding technique. More importantly, we propose a modified side‐sensitive Synthetic‐ chart, and then using overall performance measures (i.e., the extra quadratic loss, average ratio of average run‐length, and performance comparison index), we show that this new chart has a uniformly better performance than its Shewhart competitors. We also provide easy‐to‐use tables for each of the chart's design parameters to aid practical implementation. Moreover, a performance comparison with their corresponding counterparts (i.e., synthetic and RR charts) is conducted. Copyright © 2015 John Wiley & Sons, Ltd.     

Shewhart Control Charts for Monitoring Reliability with Weibull Lifetimes

In this paper, we present Shewhart‐type and S² control charts for monitoring individual or joint shifts in the scale and shape parameters of a Weibull distributed process. The advantage of this method is its ease of use and flexibility for the case where the process distribution is Weibull, although the method can be applied to any distribution. We illustrate the performance of our method through simulation and the application through the use of an actual data set. Our results indicate that and S² control charts perform well in detecting shifts in the scale and shape parameters. We also provide a guide that would enable a user to interpret out‐of‐control signals. Copyright © 2014 John Wiley & Sons, Ltd.     

Adaptive control charts for skew‐normal distribution

The standard Shewhart‐type chart, named FSS‐ chart, has been widely used to detect the mean shift of process by implementing fixed sample and sampling frequency schemes. The FSS‐ chart could be sensitive to the normality assumption and is inefficient to catch small or moderate shifts in the process mean. To monitor nonnormally distributed variables, Li et al [Commun Stat‐Theory Meth. 2014; 43(23):4908‐4924] extended the study of Tsai [Int J Reliab Qual Saf Eng. 2007; 14(1):49‐63] to provide a new skew‐normal FSS‐ (SN FSS‐ ) chart with exact control limits for the SN distribution. To enhance the sensitivity of the SN FSS‐ chart on detecting small or moderate mean shifts in the process, adaptive charts with variable sampling interval (VSI), variable sample size (VSS), and variable sample size and sampling interval (VSSI) are introduced for the SN distribution in this study. The proposed adaptive control charts include the normality adaptive charts as special cases. Simulation results show that all the proposed SN VSI‐ , SN VSS‐ , and SN VSSI‐ charts outperform the SN FSS‐ chart on detecting small or moderate shifts in the process mean. The impact of model misspecification on using the proposed adaptive charts and the sample size impact for using the FSS‐ chart to monitor the mean of SN data are also discussed. An example about single hue value in polarizer manufacturing process is used to illustrate the applications of the proposed adaptive charts.     

Monitoring the Process Mean When Standards Are Unknown: A Classic Problem Revisited

One of the most common applications in statistical process monitoring is the use of control charts to monitor a process mean. In practice, this is often performed with a Shewhart chart along with a Shewhart R (or an S) chart. Thus, two charts are typically used together, as a scheme, each using the 3‐sigma limits. Moreover, the process mean and standard deviation are often unknown and need to be estimated before monitoring can begin. We show that there are three major issues with this monitoring scheme described in most textbooks. The first issue is not accounting for the effects of parameter estimation, which is known to degrade chart performance. The second issue is the implicit assumption that the charting statistics are both normally distributed and, accordingly, using the 3‐sigma limits. The third issue is multiple charting, because two charts are used, in this scheme, at the same time. We illustrate the deleterious effects of these issues on the in‐control properties of the charting scheme and present a method for finding the correct charting constants taking proper account of these issues. Tables of the new charting constants are provided for some commonly used nominal in‐control average run length values and different sample sizes. This will aid in implementing the charting scheme correctly in practice. Examples are given along with a summary and some conclusions. Copyright © 2015 John Wiley & Sons, Ltd.     

The steady‐state behavior of the synthetic and side‐sensitive synthetic double sampling charts

The steady‐state average run length is used to measure the performance of the recently proposed synthetic double sampling chart (synthetic DS chart). The overall performance of the DS chart in signaling process mean shifts of different magnitudes does not improve when it is integrated with the conforming run length chart, except when the integrated charts are designed to offer very high protection against false alarms, and the use of large samples is prohibitive. The synthetic chart signals when a second point falls beyond the control limits, no matter whether one of them falls above the centerline and the other falls below it; with the side‐sensitive feature, the synthetic chart does not signal when they fall on opposite sides of the centerline. We also investigated the steady‐state average run length of the side‐sensitive synthetic DS chart. With the side‐sensitive feature, the overall performance of the synthetic DS chart improves, but not enough to outperform the non‐synthetic DS chart. Copyright © 2014 John Wiley & Sons, Ltd.     

A Robust Control Chart

This article studies alternative standard deviation estimators that serve as a basis to determine the control chart limits used for real‐time process monitoring (phase II). Several existing (robust) estimation methods are considered. In addition, we propose a new estimation method based on a phase I analysis, that is, the use of a control chart to identify disturbances in a data set retrospectively. The method constructs a phase I control chart derived from the trimmed mean of the sample interquartile ranges, which is used to identify out‐of‐control data. An efficient estimator, namely the mean of the sample standard deviations, is used to obtain the final standard deviation estimate from the remaining data. The estimation methods are evaluated in terms of their mean squared errors and their effects on the performance of the phase II control chart. It is shown that the newly proposed estimation method is efficient under normality and performs substantially better than standard methods when disturbances are present in phase I. Copyright © 2012 John Wiley & Sons, Ltd.     

Designing Phase I Shewhart X¯charts: Extended tables and software

Control charts play an important role in Phase I studies, which are conducted to establish process control and generate reference data for parameter estimation and calculation of prospective (Phase II) control limits. Researchers have tabulated the necessary charting constants for the normal theory–based Phase I Shewhart $\overline{X}$ chart for the process mean to achieve a desired nominal false alarm probability given the number of Phase I subgroups, m, up to 15. However, in practice, when parameters are estimated, the currently recommended number of Phase I subgroups is much larger than covered by their tables. Recognizing the need and taking advantage of some recently available software and computing resources, an extension to these tables is provided for m = 3(1)10 , 15(5)30 , 50(25)300 and n = 3 , 5 , 7 , 10. In addition to the tables, an R program is provided to calculate the charting constant, on demand, for user‐given values of nominal false alarm probability, m, and n. An appendix shows the details of how the program should be used.     

On efficient construction and evaluation of runs rules–based control chart for known and unknown parameters under different distributions

In this study, we have considered two design structures of control chart by covering the situations of known and unknown parameters, variety of probability distributions, and runs rules. The design structures are dependent on constants which generally considered hard to compute analytically. For construction of constants and also for evaluating performance of the design structures through performance measures, we have illustrated Monte Carlo simulation procedure/algorithm for researcher and practitioners. Furthermore, based on the Monte Carlo simulation procedures, we have established a program in R language to compute values of different constants and performance measures. Results illustrated that design structures for known and unknown parameters under variety of runs rules and probability distributions have outstanding performance in contrast to existing structures. Moreover, design structure for unknown parameters behaves alike the design structure for known parameters. This indicates that design structure for unknown parameters has the ability to resolve the issue of runs rules which generally occur when parameters are estimated. Besides, two real‐life examples have been included in which physicochemical characteristic of groundwater and plasticizer characteristic of petrochemical process are monitored through design structures.     

Optimal Designs of the Variable Sample Size Chart Based on Median Run Length and Expected Median Run Length

The variable sample size (VSS) chart, devoted to the detection of moderate mean shifts, has been widely investigated under the context of the average run‐length criterion. Because the shape of the run‐length distribution alters with the magnitude of the mean shifts, the average run length is a confusing measure, and the use of percentiles of the run‐length distribution is considered as more intuitive. This paper develops two optimal designs of the VSS chart, by minimizing (i) the median run length and (ii) the expected median run length for both deterministic and unknown shift sizes, respectively. The 5th and 95th percentiles are also provided in order to measure the variation in the run‐length distribution. Two VSS schemes are considered in this paper, that is, when the (i) small sample size (n_S) or (ii) large sample size (n_L) is predefined for the first subgroup (n₁). The Markov chain approach is adopted to evaluate the performance of these two VSS schemes. The comparative study reveals that improvements in the detection speed are found for these two VSS schemes without increasing the in‐control average sample size. For moderate to large mean shifts, the optimal VSS chart with n₁ = n_L significantly outperforms the optimal EWMA chart, while the former is comparable to the latter when n₁ = n_S. Copyright © 2016 John Wiley & Sons, Ltd.