The Study On The Performance Of T Type Control Charts

The performance of an X chart is usually studied under the assumption that the pro-cess standard deviation is known or standard deviation can be accurately estimated before the control chart is established.This is,of course,not always the case in practice.The study found that the X type charts are not robust against estimation errors or changes in process standard deviation.Therefore,t control chart and exponentially weighted moving average t chart(EWMA-t)are proposed.Since the t chart does not need to estimate the process standard deviation,it can overcome the impact of parameter estimation.In this paper,we present cumulative sum t chart(CUSUM-t)?double EWMA t chart(DEWMA-t)and Mixed EWMA-CUSUM t control charts(MEC-t)and determine its control limits for different parameters.We used the Monte-Carlo simulation method to measure the performance of these control charts and compared them.Numerical results show that t type control charts presented in this paper have good performance against es-timation errors or changes in process standard deviation.Especially for small and medium shift,they have better detection ability.