A. Aghaie - Associate Professor, Department of Industrial Engineering, K.N. Toosi University of Technology, Tehran, Iran.
Y. Samimi - PhD student, Department of Industrial Engineering, K.N. Toosi University of Technology, Tehran, Iran.
S. Asadzadeh - MS student, Department of Industrial Engineering, K.N.Toosi University of Technology, Tehran, Iran.

In spite of the increasing interest in monitoring multistage manufacturing processes through the recent years, this issue has been neglected to a large extent for cascade processes with dependent attribute quality characteristics. This paper is concentrated on the monitoring and diagnosing a two-stage manufacturing process where the number of defects on each inspection unit in the second stage depends on the item’s quality status in the first stage of the process. It is assumed that these dependent quality characteristics of the process follow Bernoulli and Poisson probability distributions. In addition to the evaluation of several monitoring techniques including a rather new method based on the generalized Poisson distribution, the problem of process root-cause analysis is explored especially in order to identify the peculiar stage responsible for the out-of-control conditions. It is shown that a combined approach consisting of a proportion defective control chart and an adjusted control chart for the number of defects can provide the best solution with respect to both the detectability power and correctly identification of the out-of-control stage.

multistage processes, fault diagnosis, dependent quality characteristics