Using Shewhart control charts for continuous model optimization

This paper intends to provide a view on different possibilities of using Shewhart Control Charts in order to detect sources of imprecision or inaccuracy in models for industrial processes. In the classical approach, control charts are commonly used to detect abnormal variations of a single process variable, which can be the error of a process model. The techniques explored in this work help linking this error to other variables from the model, allowing a fast detection of incorrect parameterization.