Process capability indices provide a measure of the output of an in-control process that conforms to a set of specification limits. These measures, which assume that process output is approximately normally distributed, are intended for measuring process capability for manufacturing systems. When the performance of a system results in a product that fails to fall within a given specification range, however, the product is typically scrapped or reworked, and the actual distribution that the customer perceives after inspection is truncated. In this paper, the concept of a truncated measure for three types of quality characteristics is introduced as the key to linking customer perception to process capability. Subsequently, a set of customerperceived process capability indices is presented as an extension of traditional manufacturer-based counterparts. Finally, data transformation-based process capability indices are also discussed. A comparative study and numerical example reveal considerable differences among the traditional and proposed process capability indices. It is believed that the proposed process capability index for various quality characteristics may more aptly lead to process improvement by facilitating a better understanding of the integrated effects found in engineering design problems.

Robust parameter design (RPD) aims to build product quality in the early design phase of product development by optimizing operating conditions of process parameters. A vast majority of the current RPD studies are based on an uncensored random sample from a process distribution. In reality, censoring schemes are widely implemented in lifetime testing, survival analysis, and reliability studies in which the value of a measurement is only partially known.However, there has been little work on the development of RPD when censored data are under study. To fill in the research gaps given practical needs, this paper proposes response surface-based RPD models that focus on survival times and hazard rate. Primary tools used in this paper include the Kaplan-Meier estimator, Greenwood's formula, the Cox proportional hazards regression method, and a nonlinear programming method. The experimental modeling and optimization procedures are demonstrated through a numerical example. Various response surface-based RPD optimization models are proposed, and their RPD solutions are compared.

No abstract

No abstract