In this paper, a new perturbed gamma degradation process where the measurement error depends in stochastic sense on the hidden degradation level. This new model generalizes a perturbed gamma process recently suggested in the literature, by allowing for the presence of a unit-specific random effect. The main features of the proposed model are highlighted. Model parameters are estimated, from the available perturbed measurements, by means of the maximum likelihood method. The conditional probability density functions of both the actual and the measured degradation levels, given the past noisy measurements, are computed by using a particle filtering method. Finally, a numerical application is developed on the basis of a set of real degradation data gathered via periodic inspections, where it is discussed the effect of neglecting the presence of random effect on the estimates of the cumulative distribution function of the remaining useful life of the considered degrading units. Obtained results demonstrate the affordability and prove the usefulness and effectiveness of the proposed generalization.
A Perturbed Gamma Process with Random Effect and State-Dependent Error / Castanier, Bruno; Esposito, Nicola; Giorgio, Massimiliano; Mele, Agostino. - (2020), pp. 1-8. (Intervento presentato al convegno 30th European Safety and Reliability Conference and 15th Probabilistic Safety Assessment and Management Conference tenutosi a Venice, Italy nel 1-6 November 2020).
A Perturbed Gamma Process with Random Effect and State-Dependent Error
Giorgio Massimiliano
;
2020
Abstract
In this paper, a new perturbed gamma degradation process where the measurement error depends in stochastic sense on the hidden degradation level. This new model generalizes a perturbed gamma process recently suggested in the literature, by allowing for the presence of a unit-specific random effect. The main features of the proposed model are highlighted. Model parameters are estimated, from the available perturbed measurements, by means of the maximum likelihood method. The conditional probability density functions of both the actual and the measured degradation levels, given the past noisy measurements, are computed by using a particle filtering method. Finally, a numerical application is developed on the basis of a set of real degradation data gathered via periodic inspections, where it is discussed the effect of neglecting the presence of random effect on the estimates of the cumulative distribution function of the remaining useful life of the considered degrading units. Obtained results demonstrate the affordability and prove the usefulness and effectiveness of the proposed generalization.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.