Reliability demonstration tests considering performance degradation with measurement error
C. Ruiz, H. Liao and E.A. Pohl
https://doi.org/10.19124/ima.2018.001.23
Abstract
Reliability demonstration tests (RDTs) have been widely used in engineering design to verify if a product has met a certain reliability requirement. Such tests are usually conducted and analyzed based on binomial theory for the number of failures or analysis of failure times. Unlike these traditional methods, a degradation-based RDT method is proposed in this paper. Appropriate implementation of this method will speed up reliability demonstration, especially for highly reliable products. However, a big challenge is the measurement error that cannot be avoided in degradation data collection. To incorporate the impact of measurement errors in degradation-based RDT, a random effects stochastic process model explaining both the evolution of product degradation and measurement error is proposed. Under this model, a statistical inference method based an expectation-maximization algorithm is developed to estimate the model parameters. Moreover, the optimal design of degradation-based RDT is developed to minimize the total testing cost considering both the producer’s and consumer’s risks. A numerical example is presented to illustrate the use of the proposed RDT method in practice.