摘要:

Cost control, resource availability, and/or difficulty in performing complete randomizations may dictate the necessity to run response surface experiments in a bi-randomization error control format of which the split-plot design is a special case. Response surface analysis methods along with error variance estimation techniques are discussed in the context of the resulting error structure. The efficiency of commonly used response surface designs are also examined to determine the necessity of design modifications to better accommodate the error structure.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979697

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

For multivariate statistical process control with individual observations the usually recommended procedure in the retrospective phase is the Hotelling'sT2control chart. All the observations are pooled to estimate the mean vector and covariance matrix. An out-of-control signal results if theT2value of any observation exceeds an upper control limit. We show this procedure isnoteffective in detecting a shift in the mean vector because the covariance matrix is badly estimated. Somewhat surprisingly, the signal probability decreases with increasingly severe shifts in the mean vector. We compare several alternative methods for estimating the covariance matrix and recommend a procedure analogous to the use of moving ranges in the univariate case. This procedure uses the vector differences between successive observations to estimate the in-control covariance matrix of the process. With this more robust estimate, step and ramp shifts in the mean vector are more likely to be detected in the retrospective phase.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979698

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

Statistical process control methods for monitoring processes with multivariate measurements in both the product quality variable space and the process variable space are considered. Traditional multivariate control charts based onχ2andT2statistics are shown to be very effective for detecting events when the multivariate space is not too large or ill-conditioned. Methods for detecting the variable(s) contributing to the out-of-control signal of the multivariate chart are suggested. Newer approaches based on principal component analysis and partial least squares are able to handle large ill-conditioned measurement spaces; they also provide diagnostics which can point to possible assignable causes for the event. The methods are illustrated on a simulated process of a high pressure low density polyethylene reactor, and examples of their application to a variety of industrial processes are referenced.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979699

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

In certain situations in life testing, it is more practical to record whether a unit fails in an interval instead of measuring failure times exactly. This paper considers the analysis of life data from the exponential distribution with the use of periodic monitoring. A number of different monitoring schemes are presented and compared. We show that when it is not feasible to record every measurement exactly, the use of as few as two strategically selected monitoring times can be used to analyze the data with only a moderate loss in efficiency for certain problems. The paper also presents sampling plans for a lower-tailed test for the mean of an exponential distribution using two monitoring times. Although the paper focuses on the exponential distribution, all the results also hold for a Weibull distribution with a known shape parameter.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979700

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

This paper presents two approximate methods for constructing lower confidence limits for proportion of conformance. Proportion of conformance is defined as the proportion of products with quality characteristic inside the specification limits. The performance of each method is assessed by simulation. Numerical examples are also given to illustrate the procedures.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979701

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

Two simple rules are used to develop foldovers for 2k-pfractional factorial designs. Although the rules are general, our focus is on foldovers for resolution IV designs. Foldovers that sign-reverse more than one factor can generate alias sets that cannot be obtained from a sign-reversal of a single factor. The potential alias structure can be limited by widely-used statistical software. An example illustrates the interplay of the rules and information from active alias sets in the original fractional experiment.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979702

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

When a control chart for attributes data is based on a large sample size, one often finds that the control limits are narrow, and that many observations fall outside the control limits. The narrow control limits, which are caused by the large sample size, make evident the natural variation in the probability of producing a nonconforming unit from the underlying process. This paper discusses the different types of variation that affect these control charts for attributes data and the relative merits of plotting the proportion nonconforming as attributes (P chart) versus variables (individuals chart), and it develops objective criteria for making this choice. Finally, several examples are presented, using data from a business telephone service maintenance process.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979703

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

Quick switching systems (QSSs) are used to make accept/reject decisions when inspecting a series of lots. QSSs change the number of units inspected and the acceptance criteria based on past results, in order to provide extra protection during periods of poor quality while reducing the cost of inspection during periods of good quality. This article shows how to evaluate the protection provided by QSSs during periods of both constant and changing quality, as well as how to select QSSs based on the protection they provide. QSSs are compared with single, double, chain, and variables sampling plans as well as the MIL-STD 105E switching systems. It is concluded that QSSs are superior to chain sampling plans and can dramatically reduce the amount of inspection compared to single and double sampling plans, but at the cost of increased complexity. Further, MIL-STD-105E switching systems can be replaced with far simpler QSSs that require the same number or fewer samples and that offer equivalent protection during periods of constant quality, but increased protection against isolated or initial bad quality lots. QSSs prove especially useful for processes running at low levels of defects where small sample sizes and high sensitivity to problems are desired.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979704

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

摘要:

A FORTRAN program is presented for selecting the quick switching system (QSS) with a specified acceptable quality level (AQL) and lot tolerance percent defective (LTPD) that minimizes the average sample number (ASN) at a specified point at or below the AQL. The program is limited to QSSs consisting of two single sampling plans. The maximum allowable error rate and the required rate of convergence must be specified at both the AQL and LTPD. Restrictions can be placed on the parameters of the QSS.

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979705

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor

ISSN:0022-4065    

DOI:10.1080/00224065.1996.11979707

出版商:Taylor&Francis    

年代:1996

数据来源: Taylor