ENBIS: European Network for Business and Industrial Statistics

A model of quality assurance by failure rate function

21 September 2009, 15:05 – 15:25

Abstract

Submitted by

Jan Myszewski

Authors

Jan Myszewski

Affiliation

Kozminski University

Abstract

Phenomena contributing to variability in a process are of various types. Some of them can be described by continuous probability distributions, as for instance performance of machines or properties of processed material. There are also discrete factors such as human errors or machine failures. Six sigma methodology encompasses them by assigning to the overall variability a sigma measure which can be used both to continuous as well as to discrete models when assessing performance. However it can not be used directly when assessing individual impact of specific class of factors, e.g. human errors in a continuously distributed production process.
Idea of reliability models using a failure rate has been adapted to general quality assurance case. A probabilistic model has been constructed to represent substantial phenomena of various types (continuous and discrete) that contribute to generation of nonconforming effects of human work.
The model enables intuitive graphical interpretation of causes of problems by a failure rate function and can be used to support some considerations regarding organisation and measurement of work flow during a work-shift.
It can be applied in general capability study of process when assessing individual impact of particular classes of factors to variability measures, for example to the sigma measure used in the six sigma methodology.
One of specific application areas of the model can be control of human errors. The model highlights an importance of proper work environment and a role of organisational standards as measures of prevention against human errors.

View paper

Return to programme