Wear process lifetime prediction based on parametric model applied to experimental data

Lifetime prediction of a technical system plays a significant role also with respect to the avoidance of breakdowns. The first part of this contribution is a brief review of lifetime models followed by an introduction of a new parametric lifetime model. Experimental data for the lifetime model training and evaluation are taken from a tribological system describing a wear process. The main focus of this contribution is the development of a prognosis approach for the end-of-lifetime estimation using proposed lifetime model. The impact of the number of datasets used for model training on prediction accuracy is discussed. Two cases are studied considering different number of datasets used for lifetime model training. Additionally, prediction accuracy with the approach to the end-of-lifetime (higher number of available measurements) are discussed.