Modeling wear in cylinder liners

We present and discuss a stochastic model describing the wear process of cylinder liners in a marine diesel engine. The model is based on a stochastic differential equation, and Bayesian inference is illustrated. Corrosive action and measurement error, both quite negligible, are modeled with a Wiener process whereas a jump process is used to describe the contribution of soot particles to the wear process. The model can be used to forecast the wear process and, consequently, plan condition-based maintenance activities. In the paper, we provide a critical illustration of the mathematical and computational aspects of the model. We propose a strategy that, implemented for simulated and real data, allows for stable parameter estimation and forecasts.