An enhanced nonparametric quality control chart with application related to industrial process

Area of Science:

• Industrial Engineering
• Statistical Quality Control
• Nonparametric Statistics

Background:

• Process monitoring often requires distributional assumptions, limiting applicability when data is non-normal or unknown.
• Nonparametric control charts offer a robust alternative by not assuming specific data distributions.

Purpose of the Study:

• To develop a novel nonparametric double homogeneously weighted moving average control chart.
• To monitor the location parameter of a process without assuming normality.

Main Methods:

• Development of a nonparametric control chart utilizing the Wilcoxon signed-rank statistic.
• Monte Carlo simulations to generate run-length profiles for performance evaluation.
• Comparison with existing nonparametric charts using metrics like extra quadratic loss.

Main Results:

• The proposed nonparametric control chart demonstrated superior performance compared to existing counterparts.
• Run-length profiles indicated enhanced sensitivity and efficiency.
• The chart's practical utility was validated using a dataset of automobile piston ring diameters.

Conclusions:

• The newly developed nonparametric double homogeneously weighted moving average chart is effective for process location monitoring under unknown distributions.
• It offers a statistically sound and practically applicable alternative to traditional control charts.
• The chart provides improved performance metrics, making it valuable for quality control in manufacturing.