00195 H3e245baff52db427c42443091923be

197

Strategies for Statistical Monitoring of Integral Control

compromises the statistical properties of the Shewhart, CUSUM, and EWMA to effectively perform.

In this study, two different methods are employed for SPC schemes to handle the problems associated with correlation. Lucas (1976) proposes modifying the CUSUM scheme by doubling the slack value (K) to handle in this problem. Here we consider using the values of K = Ict and 2ct. Another method explored is that proposed by Montgomery and Mastrangelo (1991) in which a moving centerline EWMA approach is employed. An equivalent approach is to use an EWMA procedurę for the predicted errors. The basie details of this approach are now outlined. First we define the EWMA of the control actions as follows

Zt = lu_t + ( 1-Ozj.!    (4)

Figurę 1. Control Actions for Process Shift of 10 Units per Period at Observation 251, PMepc/spc = 3.23.