MCEWMA control chart to detect small shift for autocorrelated water treatment process responses

Abstract

Statistical process control (SPC) chart for individual observation helps to understand the state of control, stability, and capability of a process. However, most of the typical control chart suggested for individual observation is primarily based on the assumption that the process data are independent and normally distributed. This assumption of independency and normality is generally violated in many chemical processes. Response characteristic or output may have autocorrelation (or data are time dependent), and traditional control chart becomes practically unsuitable for such situations. A two-stage time series modelling and monitoring of residual approach by using control chart or using variable control limit-based MCEWMA is generally recommended. These charts can detect small shift for autocorrelated individual responses. In this paper, the usefulness of MCEWMA chart to detect small shift for autocorrelated mineral water treatment process is verified. The ARL values for MCEWMA are also proposed based on simulation results. The case study results confirms the suitability of ECWMA and is recommended as an alternative to the two-stage time series model and residual special control chart approach.