Search results

A process improvement sampling methodology, known as process variation diagnostic tool (PROVADT), was proposed by Cox et al. (2013). The method was designed to support the objectivity of Six Sigma projects performing the measure-analyse phases of the define-measure-analyse-improve-control cycle. An issue in PROVADT is that it is unable to distinguish between measurement and product variation in the presence of a poor Gage repeatability and reproducibility (R&R) result. The purpose of this paper is to improve and address PROVADT’s sampling structure by enabling a true Gage R&R as part of its design.

This paper derives an enhanced PROVADT method by examining the theoretical sampling constraints required to perform a Gage R&R study. The original PROVADT method is then extended to fulfil these requirements. To test this enhanced approach, it was applied first to a simulated manufacturing process and then in two industry case studies.

The results in this paper demonstrates that enhanced PROVADT was able to achieve a full Gage R&R result. This required 20 additional measurements when compared to the original method, but saved up to ten additional products and 20 additional measurements being taken in future experiments if the original method failed to obtain a valid Gage R&R. These benefits were highlighted in simulation and industry case studies.

The work into the PROVADT method aims to improve the objectivity of early Six Sigma analyses of quality issues, which has documented issues.

The purpose of this paper is to examine the efficiency and objectivity of current Six Sigma practices when at the measure/analyse phase of the DMAIC quality improvement cycle.

A new method, named process variation diagnostic tool (PROVADT), demonstrates how tools from other quality disciplines can be used within the Six Sigma framework to strengthen the overall approach by means of improved objectivity and efficient selection of samples.

From a structured sample of 20 products, PROVADT was able to apply a Gage R&R and provisional process capability study fulfilling the pre-requisites of the measure and early analyse phases of the DMAIC quality improvement cycle. From the same sample, Shainin multi-vari and isoplot studies were conducted in order to further the analysis without the need of additional samples.

The method was tested in three different industrial situations. In all cases PROVADT's effectiveness was shown at driving forward a quality initiative with a relatively small number of samples. Particularly in the third case, it lead to the resolution of a long standing complex quality problem without the need for active experimentation on the process.

This work demonstrates the need to provide industry with new statistical tools which are practical and give users efficient insight into potential causes of a process problem. PROVADT makes use of data needed by quality standards and Six Sigma initiatives to fulfil their requirements but structures data collection in a novel way to gain more information.