Reliability assessment of products with Weibull lifetimes: A two-sided linked lots deferred sampling plan (T-SLLDSP)

This study proposes the Two-Sided Linked Lots Deferred Sampling Plan (T-SLLDSP), a novel statistical quality control method for assessing product reliability when lifetimes follow a Weibull distribution under time-truncated inspection, thus establishing a new, more efficient sampling framework for this specific product failure profile. The main problem addressed is the limitations of existing sampling plans, which often require unnecessarily large sample sizes, making them inefficient and costly for destructive testing scenarios, while also neglecting the consideration of producer risk and information from subsequent lots. This is important because developing a more resource-efficient plan minimizes inspection time, cost, and waste for manufacturers, while improving the accuracy of acceptance decisions to safeguard the interests of both the producer and the consumer. The study was conducted by developing the plan's mathematical framework, including the minimum optimal sample size formula and the Operating Characteristics (OC) function, followed by a simulation analysis and a comparative performance test using a real-life dataset of oil palm milling machine ball-bearing failures. It was discovered that the Weibull distribution's failure rate decreases as the mean life ratio increases, and the T-SLLDSP consistently requires a smaller sample size than existing single, double, or chain sampling plans. These findings mean the T-SLLDSP provides a superior balance between the statistical risks and the inspection cost. Ultimately, these findings are useful and important because they offer manufacturers a robust, economical, and risk-managed approach to quality assurance, especially beneficial in industries where high testing costs and destructive sampling are critical constraints.