ABSTRACT
The Gaussian process (GP) approach is one such data-driven approach that can be used for fatigue damage prediction of complex geometry and under complex loading patterns. This chapter discusses the use of GP for fatigue life prediction of an Al-2024 cruciform structure under Fighter Aircraft Loading STAndard For Fatigue and random loading. The fracture mechanics–based Paris law is widely used for fatigue life prediction of structural components. The constant cycle fatigue tests were performed for 20–30 kilocycles to obtain an initial crack of 1–3 mm. For the demonstration of the prognosis algorithm, crack lengths were estimated at discrete instances. A Bayesian GP predictive model has been developed to forecast damage under random and flight profile fatigue loading. The GP model is based on high-dimensional kernel transformation that can perform nonlinear pattern reorganization. The framework is used to predict the future fatigue damage states of aluminum cruciform specimen under biaxial loading.
