A stochastic process-based degradation modeling framework considering measurement errors: a perspective of dual non-Gaussian assumptions.

AbstractThe stochastic processes are a natural choice for describing the randomness in degradation processes caused by inherent randomness and environmental factors. This article proposes a new modified skew-normal distribution to capture measurement uncertainty, and then establishes a stochastic process-based degradation modeling framework, in which both degradation increments and measurement errors do not follow the Gaussian distributions. Taking the IG process as an example, the basic reliability indicators and the alarm probabilities caused by measurement errors are derived. In addition, a multi-stage parameter estimation algorithm based on moderate particle sizes and comparative tolerances is developed for this stochastic process-based degradation model under dual non-Gaussian assumptions. Finally, the effectiveness and advantages of the proposed model along with the parameter estimation algorithm are demonstrated by a simulation study and a case application, and it is particularly pointed out that the relative size of measurement errors has a significant impact on the precision of degradation reliability assessment. [ABSTRACT FROM AUTHOR]