EWMA and DEWMA repetitive control charts under non-normal processes.

Area of Science:

• Industrial Engineering
• Statistical Process Control
• Quality Management

Background:

• Traditional control charts often assume normal process distributions, limiting their effectiveness for non-normal data.
• Monitoring process shifts is crucial for maintaining product quality and operational efficiency.
• Existing methods may not be optimal for detecting early shifts in non-normal processes.

Purpose of the Study:

• To develop and evaluate a repetitive sampling method for constructing EWMA and DEWMA control charts.
• To enhance the ability to monitor shifts in non-normal processes using various distributions.
• To compare the performance of proposed charts against existing methods using simulation and real data.

Main Methods:

• Utilized a repetitive sampling strategy for EWMA and DEWMA control charts.
• Employed t-distribution (symmetric), gamma distribution (positively skewed), and Weibull distribution (negatively skewed) for non-normal data.
• Conducted Monte Carlo simulations and analyzed average run length (ARL) to assess chart performance.
• Validated the proposed charts on two real-world datasets.

Main Results:

• The proposed control charts demonstrated superior ability in detecting process shifts earlier compared to existing charts.
• The repetitive sampling method proved effective across different non-normal distribution types.
• Simulations confirmed the practical utility and enhanced sensitivity of the new charts.

Conclusions:

• The developed repetitive sampling method for EWMA and DEWMA control charts offers a robust solution for monitoring non-normal processes.
• These charts provide improved sensitivity for early detection of process shifts, contributing to better quality control.
• The findings suggest practical applicability in industries dealing with non-normal process data.