Testing and Data Analysis Methods for the Durability and Reliability Assessment of the XNO_x SCR System

摘要

To efficiently validate the durability and reliability performance of the new XNO_x selective catalytic reduction (SCR) system, an emission control product for reducing nitrogen oxides (NO_x), several testing and associated data analysis methods have been adopted and utilized over the past years. In this paper, two fundamental and widely used methods, i.e. the test-to-failure method (for complete and right censored data) and the Binomial test method, which is based on the Bernoulli trial and the Binomial probabilistic distribution, are described in detail. The concept of RxxCyy (e.g. R90C90 indicates 90% in reliability and 90% in confidence) is the criterion currently used in both the test-to-failure and the Binomial test method for product validation. The most recent developments made by the authors of this paper are also provided with emphasis on sample size reduction, testing time/cost savings as well as accuracy/confidence increase in probability-based predictive tools. Three specific techniques, which demonstrate significant advantages in achieving these goals, are highlighted: (1) the Bayesian statistics approaches for complete data, (2) accelerated testing approaches for Binomial test method, and (3) the Imprecise Probability approach to address the uncertainties in product design. Some of these methods have been implemented into the validation procedures for new products including the XNO_x SCR system, in which the time-to-market and budget constraint are critical. The rest of the methods are being verified and validated for possible further product implementation.