Control charts using half-normal and half-exponential power distributions using repetitive sampling

Area of Science:

• Industrial Engineering
• Statistical Quality Control
• Reliability Engineering

Background:

• Manufacturing processes require robust methods for monitoring defective items.
• Traditional control charts may not fully capture lifetime data characteristics.
• Time-truncated life tests offer a framework for analyzing data with limited observation periods.

Purpose of the Study:

• To develop and evaluate attribute control charts (ACC) for monitoring defective items in manufacturing.
• To tailor these charts using time-truncated life tests (TTLT) for specific lifetime distributions.
• To assess the performance of proposed charts under repetitive sampling schemes (RSS).

Main Methods:

• Development of ACCs based on the half-normal distribution (HND) and half-exponential power distribution (HEPD).
• Application of time-truncated life tests (TTLT) and repetitive sampling schemes (RSS).
• Performance evaluation using average run length (ARL) calculations for in-control and out-of-control scenarios, considering parameter shifts.

Main Results:

• The HEPD-based ACC significantly outperforms HND-based and Exponential distribution (ED)-based ACCs in detecting process shifts, evidenced by lower ARL values.
• The HND-based ACC also shows improved effectiveness compared to the ED-based ACC under TTLT and RSS.
• Simulation testing and real-life implementation confirm the practical applicability of the proposed control charts.

Conclusions:

• The proposed HEPD-based attribute control chart is highly effective for monitoring manufacturing quality.
• The HND-based ACC offers a valuable alternative for quality control applications.
• These advanced control charts enhance defect detection and process monitoring in manufacturing settings.