The performance of control charts using transfer function

The procedure to accommodate continuous data with autocorrelation is to use control charts residuals which are adjusted by time series methods, and when we analyze discrete data with autocorrelation is analyzed, the Poisson distribution is often employed. This paper shows the application of statistic quality control (SQC) to the sand foundry in the Industry FUNDIMISA, in Santo Angelo (RS/Brasil). The variable plasticity of the sand was analyzed before the elaboration of the broken casts, and verified its relation with the non conformity of those broken casts. The objective of this paper is to investigate the performance and adequacy of the traditional use of control charts with autocorrelated data, and to motivate the use of time series methodologies to fit the correlated continuous variable; and the use of Poisson regression models to fit the correlated discrete variable, and to discuss the use of transfer function models in a productive process.