Data-driven method for real-time prediction and uncertainty quantification of fatigue failure under stochastic loading using artificial neural networks and Gaussian process regression

Various engineering systems such as naval and aerial vehicles, offshore structures, mechanical components of motorized systems, are exposed to fatigue failures due stochastic loadings. Methods for early failure prediction essential engineering, military, civil applications. In addition the time (TtF), uncertainty quantification (UQ) is major importance real-time decision-making purposes. Usually, domain or frequency methods used prediction, rainflow counting Miner’s rule Dirlik’s method. However, those suffer from over-simplistic modeling inaccurate predictions under During last years, several data-driven models were suggested offline failure. most them not capable both accurate UQ. current work, a probabilistic model introduced. A hybrid architecture fully connected artificial neural network (FC-ANN) Gaussian process regression (GPR) proposed ensure enhanced predictive abilities simultaneous UQ predicted TtF. The performances validated using synthetic experimental data. This novel method extends forecasting capabilities existing time-domain machine learning-based prediction. It paves way towards development preventive system that provides safety operational instructions insights structural health monitoring (SHM) purposes, allowing prevention environmental damage, loss human lives.