Traditionally, prognostics has been performed by continuous monitoring of quantities that are empirically assumed to sum up the evolution of the observed phenomenon. Only recently, physics-based model simulations have been considered as a valid help to understand the evolution of the chemical-physical processes leading to failure. Nevertheless, physics-based models have always been paired with data driven models and that has led to the development of heavy hybrid approaches. It is opinion of the authors that in a wide range of situations, with the current knowledge of physics, it is possible to model processes and their evolution based only on the physics model simulations, using what's called “propagation of uncertainty technique” on every aspect, from components to system behavior. The authors developed such a technique valid for both Time-Independent and Time-Dependent phenomena. This approach allows probabilistic simulations to be carried out. The proposed methodology simultaneously uses the rated values of the involved parameter in the considered process and their uncertainties as inputs. As outputs it provides the expected values of the process with their standard deviations (std) and their probability distributions. The authors developed also the software tool RELYSOFT which allows to directly supply rated values and their uncertainties as inputs. The simulations can be performed by means of analytical equations in definite form, differential equations or by Finite Element approach. The Finite Element simulation will be the subject of a later work. In case of simulations by means of analytical equations in definite and differential form, a new pseudo algebra has been developed in order to perform assessments. Two examples are reported concerning aero-spatial field: a Time-Independent and a Time-Dependent.
Prognostics via physics-based probabilistic simulation approaches
2016-06-01
3672300 byte
Conference paper
Electronic Resource
English
Prognostics-Based Probabilistic Fatigue Life Prediction
British Library Conference Proceedings | 2003
|Prognostics-Based Probabilistic Fatigue Life Prediction
AIAA | 2003
|Probabilistic model based algorithms for prognostics
IEEE | 2006
|