Although process capability analysis for non-normal data has been used lately, there are only few works on this subject. This article presents the major methods for process capability analysis with non-normal data. Tree methods are presented, which were proposed by Clements in 1989, Pearn and (1997), and Chen and Ding (2001). The first two methods are similar in many aspects, but each one presents novelties in the computation of the capability indexes. They present indexes similar to the traditional ones, which assume normality. The method by Chen and Ding (2001) brings as novelty the use of an index to estimate the number of nonconforming items produced in the process. After the review of these methods, they are compared by means of their application in two cases. In the first case, the methods by Clements (1989) and by Pearn and Chen (1997) were analyzed to identify the best for the non-normal process capability determination. In the second case, all three methods were applied in a metallurgical industry data set. In both cases, the free software R (version 2.2) was used for the statistical analysis.
Capability index; Non-normal data; Capability
