Time filter

Source Type

Noorossana R.,Islamic Azad UniversitySouth Tehran Branch | Fatemi S.A.,Islamic Azad UniversitySouth Tehran Branch | Zerehsaz Y.,Islamic Azad UniversitySouth Tehran Branch
International Journal of Advanced Manufacturing Technology | Year: 2014

In many practical situations, quality of a process or product can well be characterized by a relation between a response variable and one or more explanatory variables. This functional relationship which is commonly referred to as profile has been addressed by many researchers for the case of fixed explanatory variable. However, there are some situations where this assumption may not hold. In this paper, we try to investigate the effect of random explanatory variable on phase II monitoring of simple linear profiles. Average run length criterion is used to evaluate both the in-control and out-of-control performances of the control schemes commonly considered for phase II profile monitoring. The charting performance of the EWMA-3 control scheme has been improved using a novel technique. © 2014, Springer-Verlag London.


Noorossana R.,Islamic Azad UniversitySouth Tehran Branch | Zerehsaz Y.,Islamic Azad UniversitySouth Tehran Branch
International Journal of Advanced Manufacturing Technology | Year: 2015

In this paper, we study the effect of classical additive measurement error model on the most commonly used control charts for monitoring simple linear profiles with random explanatory variable. We showed that the in-control and out-of-control performances of these methods are significantly affected when measurement error is present in the explanatory variable. The average run length criterion is applied to represent the performance of the methods. For instance, we can clearly see that even a small amount of measurement error introduced in the explanatory variable of a profile can increase the false alarm rate about 55 %. With the same negligible amount of error, the out-of-control average run length of the exponentially weighted moving average chart (EWMA)-3 method in detecting a moderate-size shift in the slope parameter increases 15 %. Two different strategies on the basis of control limit modification have been proposed to compensate for the undesired effect of measurement error. Results indicate that the proposed strategies substantially offset the poor performances of the control charts resulted from the presence of measurement error. © 2015, Springer-Verlag London.

Loading Islamic Azad UniversitySouth Tehran Branch collaborators
Loading Islamic Azad UniversitySouth Tehran Branch collaborators