Online Bayesian prediction of remaining useful life for gamma degradation process under conjugate priors

TL;DR

This paper introduces a conjugate prior for the gamma process, enabling efficient online prediction of remaining useful life in degradation modeling, with algorithms validated through simulations and real case studies.

Contribution

It develops a conjugate prior for the gamma process and designs algorithms for efficient online RUL prediction, addressing complex parameter inference challenges.

Findings

Algorithms show high computational efficiency and estimation accuracy.

The online prediction method performs well in real case studies.

Conjugate prior simplifies Bayesian inference for gamma degradation models.

Abstract

Gamma process has been extensively used to model monotone degradation data. Statistical inference for the gamma process is difficult due to the complex parameter structure involved in the likelihood function. In this paper, we derive a conjugate prior for the homogeneous gamma process, and some properties of the prior distribution are explored. Three algorithms (Gibbs sampling, discrete grid sampling, and sampling importance resampling) are well designed to generate posterior samples of the model parameters, which can greatly lessen the challenge of posterior inference. Simulation studies show that the proposed algorithms have high computational efficiency and estimation precision. The conjugate prior is then extended to the case of the gamma process with heterogeneous effects. With this conjugate structure, the posterior distribution of the parameters can be updated recursively, and an efficient online algorithm is developed to predict remaining useful life of multiple systems. The effectiveness of the proposed online algorithm is illustrated by two real cases.