Presenting new control approaches for monitoring quality characteristics with Weibull distribution under type 2 censorship in a two-step process
Volume 6, Issue 1, Spring 2016, Pages 9-20
Shervin Asadzadeh, Fatemeh Kiadaliry
Abstract In this paper, control diagrams are proposed to monitor the scale parameter of reliability data with Weibull distribution in the presence of type 2 censorship in cascading processes. A cumulative sum control diagram and a probability range control diagram are intended to detect decreasing shifts in the mean of the qualitative characteristic with the reliability fluid. The proposed control approaches are based on the distribution of the smallest limit value converted from the Weibull distribution to take into account the cascading property that is the main feature of multistage processes. Then, to evaluate the proposed control charts, a simulation is performed in which the comparison index of control charts is the average length of the sequence. An additional quadratic loss index has also been used to compare the ability to detect proposed control charts. In addition, sensitivity analysis has been studied to investigate the effect of the number of failures on the performance of the proposed control diagrams and the robustness of the monitoring approaches versus shifts in the previous stage of the process. Finally, to illustrate the performance of control charts, a case study from a glass bottle factory is reviewed. The results show the superiority of the cumulative control chart over the control chart with probability limits.