S
What pre-processing method(s) would you recommend to calculate the capability for a data set that follows an ARMA(2,2) process? Since there is significant positive correlation the variance will be underestimated and therefore will produce control limits that are too small. This data comes from a catalyst temperature with an upper limit only. The temperature of the catalytic bed is determined by the reaction rate of SO2 -> S03. The concentration of SO2 is diluted to reduce the reaction and hence the temperature. The dilution is via a control loop, hence the ARMA behavior.
Thanks
Thanks