MaxD: an attribute control chart to monitor a bivariate process mean

The aim of this paper is to propose a new attribute control chart, namely M a x D , to monitor a bivariate process mean. No measurement is taken but using a gauge ring each sampled unit is classified as failed/disapproved or not in each dimension. If the number of disapproved units of any dimension is larger than a control limit, a signal of out-of-control is triggered. The development of the control chart is based on a bivariate binomial distribution. The results show that the current proposal has a good performance and it is a feasible alternative to T 2 control chart mainly if the cost of the measurement of the quantitative characteristics is expensive or demands much time or the sampled units are damaged and have to be discarded in case of destructive experiments.