Robust Holt-Winter Based Control Chart for Monitoring Autocorrelated Simple Linear Profiles With Contaminated Data

Profile monitoring is a useful technique in statistical process control used when the product or process quality is represented by a function over a time period. This function represents the relationship between a response variable and one or more explanatory variables. Most existing control charts for monitoring profiles are based on the assumption that the observations within each profile are independent of each other which is often violated in practice. Sometimes there are one or more outliers in each profile, which leads to poor statistical performance of the control chart. This paper focuses on Phase II monitoring of a simple linear profile with autocorrelation within profile data in the presence of outliers. In this paper, we propose a new combined control chart based on the robust Holt-Winter model to decrease the effect of outliers. We first evaluate the effect of outliers on the performance of the proposed combined control chart. Then, we apply robust Holt-Winter and design a robust combined control chart to overcome the effect of outliers. The performance of the proposed robust Holt-Winter control chart is evaluated through extensive simulation studies. The results show that the proposed robust control chart performs well.