Sampling plans for vehicle component reliability verification

Today's vehicle customers' demand for better quality and reliability is increasing. A typical reliability requirement is that of a lower bound on product reliability at a pre-specified design life target (called the ‘bogey’) with specified confidence. For the purpose of reliability verification, one way to verify conformance to the requirements is to perform a test and terminate it after running for a certain period of time, cycles or mileage and to make a decision according to the test criterion. The approach presented in this paper specifically addresses the means to design tests that satisfy these important requirements and objectives. All the mathematical derivations are primarily based on a component life distribution that is assumed to follow a Weibull distribution with an assumed value of the shape parameter. By the proposed approach, only the GEM table (derived from the exponential case) is required. The test engineer has the option of trading off between the quantity of samples and the test bogey ratio (the test length relative to the bogey) so as to meet the specific objectives under the constraints.